summary of how the DMAIC framework was

summary of how the DMAIC framework was applied. Organize as follows: Define: In at least one detailed paragraph, summarize the define phase of this study, including the opportunity for improvement, input, customer, current process, and the output of the process. Measure: In at least one detailed paragraph, discuss the measure phase of this study. What baseline measures were taken? What is the process cycle efficiency (PCE) based on Figure 1? Analyze: In at least one detailed paragraph, indicate various tools that were utilized in the analysis phase of this study. What are the Poka-Yoke measures being considered? Improve: What improvements are being suggested, and how will they be implemented? What is the PCE measure after the implementation (use Figure 5)? What are the highlights of some of the cost savings that were achieved? What additional improvements can suggest? at least two detailed paragraphs. Control: This case study does not include many details about the control phase. If you were in charge, in addition to the control information given here, what else would you do to manage the quality of RFID in clinical care processes going forward? Does your facility use RFID technology? Can anticipate what may go wrong with this technology, and how would detect fand mitigate it? What statistical and graphical tools would use? at least two detailed paragraphs. Define Phase The major purpose of the Define phase is to identify characteristics of the system that are critical to quality to the ‘customer’ (CTQ). The study involved examining the system constituting a regular scheduled hospital outpatient surgery. The objective of this phase was to determine what creates a) errors and b) bottlenecks or restrictions on the availability of people (e.g., doctors, nurses, technical staff, and administrative staff), places (e.g., operating rooms, recovery rooms, x-ray rooms, and hospital beds) and things (e.g., equipment, tools, and supplies) as well as identify the various costs associated with the outpatient surgery process. The outpatient surgery center that was modeled was based on a major urban hospital in Minnesota which performs a dozen or so different kinds of procedures on approximately 7,800 patients per year. Typical surgeries include: cataract, eyelid, hernia, veins, nerve, podiatry, ACL, shoulders, cosmetic-breast augmentation, liposuction, plastic surgical procedures, and orthopedic hand surgeries. Gross revenues for 2009 were $50 million. The outpatient surgery center was supported by 12 doctors (general primary care physicians) for pre-ops, 12 surgeons, 22 nurses and 7 technicians. The facility has 10 operating rooms, 6 exam rooms and 10 recovery beds. 8 The research evaluated was whether implementing RFID can improve effectiveness and efficiency in that process by: • Improving traceability and visibility of services and the resources required to produce those services • Improving the coordination of the resources required to produce the services • Decreasing cycle time of tasks involved in producing the services • Improving infrastructure and cooperation among and between resources • Reducing the likelihood of errors within the processes through the use of poka yokes. • Inspiring confidence of (reduction of the anxiety of ) patients by improving patient safety • Reducing chances for infection caused by factors that may be reduced through the use of RFID: o Failure to follow sterilization protocol o Lack of cleaning/washing o Deficient operating room procedures • Reducing total costs of outpatient surgery process stream. Measure Phase The first step of this phase is to determine the metrics by which to measure the existing process. It is only by having data on the performance of the existing system that a model of the improved state can be created and a plan developed to achieve it. The metrics chosen for the study were: a) Cycle times of individual tasks within the process b) Overall throughput time of the process c) Costs of providing the procedures d) Utilization of the resources e) Reduction in errors (first time quality) As part of the measure phase, a current-state value stream map (csVSM), shown in Figure 1, was created to fully describe and define the processes on which the study intended to focus. [insert Figure 1 about here] This csVSM analyzed the basic value chain that exists in the hospital system. The map indicated there were significant areas that could be improved, specifically with regard to the need to eliminate the non-value added time that was spent on locating equipment and supplies at various process stages. It also identified problems that could arise from complications created by improper procedures and / or sterilization which occurred prior to, during, and between surgeries. These two areas were specifically targeted for process improvement using RFID technology. Analyze Phase The analyze phase of DMAIC focuses on the causes of variation and errors. It then looks to identify the root cause of those issues. It does this by promoting a thorough knowledge of what the process is, what it currently does, what it should do, what its capabilities are, and how the processes should be controlled. This study used several six sigma tools to accomplish that. To analyze the variation found in the processes and the effects of implementing RFID within those processes to reduce that variation, the following tools were used; process mapping and a discrete event simulation. Process mapping simply involves diagramming and focusing on the individual tasks needed to transform inputs into a particular output, the service. It is also the bottom half of the value stream map, already discussed. The next step was to analyze and determine how to incorporate RFID in order to improve the system. RFID Implementation The RFID implementation would include several aspects of the technology including the following: 1) Each patient wristband would incorporate a passive RFID chip (passive tags are approximately $.10 each). 2) Each employee nametag would incorporate a passive RFID chip. 3) All hospital equipment would bear an identification sticker that contained a passive RFID chip. This includes technical medical equipment (crash carts, portable x-ray, monitors) and housekeeping medical equipment (beds, wheelchairs, gurneys). RFID tags that can be sterilized are available. 4) Bulk hospital supplies and drugs can be labeled with active RFID tags (which can maintain a record of environmental temperature, tampering, etc.) while individual containers or packs would have passive tags. Active tags (approximately $50 each) are reusable and reprogrammable. 5) Separate RFID tag readers will be assumed to be unnecessary as most hospitals are able to use their existing wireless area local networks (WLAN and often referred to as Wi-Fi) to act as RFID readers. Particular areas not covered by an access point (a hotspot) may require a separate reader. For the costs in the table, it was assumed each operating room would need a separate reader. The approximate costs of this system are broken out in Table 1, Specific RFID Implementation Costs. Costs of the RFID system can vary tremendously based on the equipment, software and infrastructure. Costs for one hospital were as low as $35,000 for software and $50 for each active RFID tag (interview and correspondence with Chris Wassel, RFID Program Manager of the RFID Center of Excellence, Penn State Erie, The Behrend College, in Erie, PA). That hospital did not need readers as they used an existing internal Wi-Fi structure. Due to the wide-spread use of wide-area wireless networks, this study assumed the same. One way to look at the operational benefits, and operational assumptions, of that implementation is to describe it in terms of the errors that can occur in the processes (as identified in the current value stream map) and methods of preventing those mistakes from occurring. A term borrowed from business that describes tools and methods implemented to prevent mistakes from occurring is poke-yoke. Poke[1]yoke is a Japanese term that, when translated into English, means mistake-proofing. It came to prominence as American auto manufacturers began to adopt the ‘lean’ production techniques made famous by the Toyota Corporation. In another study, one of the authors identified where various poka-yokes can be strategically applied in a hospital setting (Kumar and Steinebach, 2006). Described below are the possible failures identified in that article and a number of mistake-avoidance measures (poka-yokes) that incorporate RFID, including how these identification tags can be tactically implemented in an outpatient surgery process. Poka-yoke: nationwide patient number. Every patient in the U.S. with a unique nationwide patient number (like a social security number) imbedded in an RFID tag worn on the wristband of the patient, would prevent hospital staff members, as well as pharmacists, from mixing-up patients or patient records. With readers located at strategically situated positions throughout the hospital, along with the location capabilities provided by the existing wireless infrastructure in most hospitals, the patient’s identification and exact location are constantly monitored in real-time. Failure II: patient mix-up (continued). Poka-yoke: wrist band, including nationwide patient number, name and birth date. When the patient arrives, the administrative person or nurse who works at the front desk of a hospital admits the patient. They should wristband with an RFID chip that contains the patient’s national patient number, birth date, and name along with a synopsis of the patient’s medical history and reason for visit. This band is then put on the patient’s wrist. Failure III: patient mix-up (continued). Inaccurate information on patient record/mix-up of patient records. Poka-yoke: wristband, including national patient number, name and birth date connected to a nationwide online database. The use of a wristband containing an RFID chip alleviates the need to physically scan the patient’s wristband each time identification is required. When a patient enters a room, either a specific reader or the wireless infrastructure recognizes the data on the chip and instantaneously brings up the correct patient’s information on the computer in that room. Failure IV: miscommunication between physicians and nurses. Poka-yoke: diagnosis and notifications must be completed through the online record. Doctors and nurses will be required to complete everything in the online patient record (diagnose, notifications, comments, medications given, medical problems, allergies, etc.). When the patient goes to the next step, if he has to be scanned again, then the updated information will appear. As physical scanning is no longer required, this step is significantly simplified, and therefore speeded up with reduced errors, by the use of RFID. The correct patient’s file is automatically displayed for updating as soon as the patient enters a new room or area. While not completely eliminating miscommunication, the practice should reduce it. (Kumar and Bauer, 2009) Failure V: patient is given the wrong medication or medication dosage. Poka-yoke: online medication ordering system. To avoid medication errors, ordering the patient’s drugs online would be better than ordering the paper prescriptions. To avoid any source of error, a physician would type them into an online form. The system then checks if the prescribed drug exists. Since the medication has already been ordered online for post-operative medications, the hospital would issue a pharmaceutical ID with an RFID tag imbedded in it that would be required for purchase of needed medications. It would be this ID that is read at the participating pharmacy. In addition, under developing mandates from the government, drug containers will have RFID tags as well; though the initial purpose of the mandates requiring these tags is to facilitate documentation of the pedigree of the drug (reducing the infiltration of counterfeit drugs). Failure VI: knowledge errors/wrong diagnosis/wrong medication/wrong medication dose. Poka-yoke: continuing obligatory training/education for doctors and nurses. Incorrect diagnosis is capable of triggering not only one, but several medical errors as every succeeding physician or nurse relies on the primary care-doctor’s diagnosis. This may occur with lack of familiarity of new methods of surgery and anesthesia by physicians and nurses. Data on physician’s and nurse’s continuing education would be maintained on a national database. As each hospital employee carries an identification card containing an RFID tag, the attending employee’s credentials are recognized when they enter a treatment area. If credentials are not current, a warning is noted on the local computer screen. Failure VII: lack of experience. Poka-yoke: training on the job. Data on physician’s and nurse’s continuing education would be maintained on a national database. As each hospital employee carries an identification card containing an RFID tag, the attending employee’s credentials are recognized when they enter a treatment area. If credentials are not current, a warning is noted on the local computer screen. Failure VIII: fatigue. Poka-yoke: working limit. Medical errors caused by physician or nurse fatigue can be reduced through the implementation of maximum working limits. Shift hours are noted when employees check in at the beginning of their shift. Once again, when shift hours near a maximum, a warning is noted on the screen in their treatment area. Once the maximum number of hours is reached, a different warning is issued when the employees RFID tag is read in a treatment area. Failure IX: negligence. Poka-yoke: checklists. Negligence can be reduced through implementing checklists. In hospital areas (i.e. in the operating room) that have high adverse event probabilities, if equipment is RFID tagged, then the trays or carts themselves can indicate what equipment is in place and what equipment is missing, both simplifying and speeding up the process of reconciling lists prior to, during, and following surgery. In this study, we examined several possible ways to poke-yoke, or prevent errors using RFID technology but focused on two additional areas of failure: 1) inability to locate specific supplies/equipment needed to perform surgery at a specific place and time (as seen in Correa et al (2007), Fredendall, et.al (2009, Tu et al (2009) and Davis (2004) mentioned in related work) and 2) Complications arising from not following procedure and improper sterilization (as mentioned in Kumar and Steinebach (2006), also mentioend above). Each of these failures lower the effectiveness and efficiency of the system because they increase the variation in the processes, as defined in the Introduction section. It should be noted that only that portion of the non-value-added time that can be saved by the use of RFID was considered. As long as humans are involved in the tasks, there will always be errors and complications that RFID cannot eliminate. What this study examined was the effect of implementing poka yokes using RFID, particularly those poka yokes relating to reducing the possibility of patient misidentification, missing or lost supplies and equipment, and complications created by fatigue, sterilization issues, or training. The effect of implementing them is modeled through time savings and reduced reentry of patients into the system from postoperative infections. This was accomplished in two stages. First, using data from hospitals and secondary sources a model was created to simulate the process without RFID. Then, using data from secondary sources on the individual improvements in time and costs created by developing each of the separate poka yokes using RFID, a simulation model was created that aggregated those benefits through the system. RFID would help to poke-yoke these failures in the following ways; locating supplies and equipment is facilitated as the hospitals wireless network system can, in real time, determine the location of any supply or piece of equipment bearing an RFID tag by the unique signal it returns when queried by the system. Its signal is picked up by the wireless system and its location displayed on any computer screen. By reducing the search for either items or space, non-value added time (and the cost associated with that time) is also reduced as is the variation created by the addition of the non-valued added time. Reduction of complications caused by post-operative infections would take place as the tags on employees would determine whether that employee had been in proximity of sterilization areas (sinks), display the appropriate procedures for a particular patient (after reading the patient’s RFID wrist tag) in a particular operating room and require the appropriate employees to electronically sign off that they have read that procedure. The reduction of post-operative infections also reduces variation created by randomly reintroducing patients back into the system. While implementing RFID is daunting, and is usually done in steps, the proposed system in the study assumes that the technology is being adopted by the hospital system as a whole. It is only in this manner that the full benefits of RFID can be realized. To determine the outcomes of implementing these poke-yokes, or error-proofing ideas, the study needed to analyze their effect on the current system. Experimenting on an actual system, however, without knowing the consequences can cause unneeded disruptions and costs. Simulation provides a way to understand implementation ramifications without disrupting the actual system. RFID vs. Bar-codes Many hospitals already use bar-coding for identification purposes. This brings up the of what are the benefits of incurring the additional expense of implementing RFID? While bar-coding has the advantage of lower cost, it also has two distinct disadvantages. 1) In order to input or read a barcode, an individual must be able to physically see and then scan that visible code with another physical piece of equipment, a reader. This makes its application in locating missing items essentially the same as having no system at all. With RFID, a tag can be scanned by the existing wireless system within the facility (or by a stationary tag reader as the tag passes near to it) with no additional effort by another party. 2) A barcode holds a single piece of information. RFID tags can hold multiple pieces of information and can, in the case of active tags, record information, such as ambient temperature, movement, and tampering that occurs on or within the item tagged. In the case of the types of errors being studied here, bar-codes would not be able to prevent the failures. Simulation Model The models developed by Kumar, et.al. (2009), and Elbeyli and Krishnan (2000) provided a starting point by which to analyze the systems under study. The first step was to define the tasks in the process. To illustrate this, a basic outpatient surgical process flow map, shown in Figure 2, was created. [insert Figure 2 about here] The dependent variables of the model are processing time (throughput time), costs of resources, and resource utilization with the independent variable being the difference in tasks arising from the implementation of RFID (and described earlier). Explanation of RFID Model Processes Though based on data from hospitals, derived from both secondary sources and interviews, the models are designed to represent a generic hospital outpatient surgery system. It assumes that patients are scheduled for an appointment prior to their hospital procedures (non-emergency). Times for the various tasks within this process are randomly selected by the model from time distributions in order to represent the stochastic nature of such a system. The parameters of the time distributions for the tasks, and the number of resources available for each task were based mainly on literature from documented empirical studies of similar systems. These sources are noted as the model is described below. If not so noted, times were estimated based on interviews with hospital administrators and staff. In reality, of course, all the parameters can vary tremendously from situation to situation. The focus in this study, however, is on the difference in overall time and costs between the scenario without RFID and the scenario with RFID. Since all other times and costs, those not affected by RFID, are held constant between the two scenarios, it is the relative difference between the models that emerges as the crucial measure for analysis. Base Model without RFID The base model begins with arrivals of scheduled patient entities at the hospital every 90 minutes. The total number of arrivals on any given day is based on a triangular distribution of T 25, 30, 45. This distribution is based on the demographics of the hospital described in the Define phase through interviews with hospital administrators. Law (2007) notes that in the absence of sufficient actual data, as is the case here, a triangular distribution is appropriate for modeling arrival rates and as an alternative to the Poisson normally used. As soon as patients arrive, they are processed through Scheduled Admittance. This process requires a resource called Staff Pack which has a capacity of 5. The process requires an amount of time selected by the model from a triangular distribution with a minimum of 5 minutes, a maximum of 30 and a most likely of 15 (T 5, 15, 30). Following admittance, each patient entity is assigned certain qualities including a time attribute beginning at zero to track their throughput time within the model. The next process is the outpatient preoperative examination. This process requires the use of four resources, each of which must be present in order for the process to proceed; supplies, exam room, doctor and nurse. Total capacity for the supplies resource is infinite. Capacities for the room, doctor, and nurse are 6, 12 and 22 respectively. Time required for this process is based on a triangular distribution with a minimum of 10 minutes, a maximum of 25 and a most likely of 20 (T10, 20, 25). At this point, there is a process within the model entitled Find Supply that assumes there is some effort required to locate the necessary supplies to perform the surgical process based on studies discussed in the section on related work. The time required for this process is based on a triangular distribution with a minimum of .5 minutes, a maximum of 15 and a most likely of 5 (T .5, 5, 15). Any time the resource supplies is required, there is a corresponding Find Supply process that follows the process in which the resource is used. The next step in the model is a process to represent pre-operative tests which could include such things as blood tests and x-rays. This process requires four resources; supplies, technician set, test room set and lab equipment set. The capacities for the latter three are seven, two, and two respectively. The time needed to perform these tests is based on a triangular distribution with a minimum of 10 minutes, a maximum of 30 and a most likely of 20 minutes (T10, 20, 30). This distribution comes from a study done by Ahmed and Alkhamis (2009). Preceding the tests is a decision point that allows for the possibility that patients do not need any preoperative tests, but it was assumed, for this model, that at least some tests (e.g., blood tests) are always needed prior to surgery. Once again, there is a “Find Supply” process that accompanies the testing process. This time, it requires the use of the technician resource and takes T .5, 10, 20 minutes. The next process in the system is to interpret the patients’ tests. This requires a doctor pack resource and takes T 10, 20, 25 minutes. Following this is a decision point that asks whether additional tests are required based on the interpretations of the initial tests. The assumption is that additional tests are required in 5% of the cases. These 5% are returned to the initial testing stage. The balance of patients move on to the next process; Outpatient Treatment. This process requires six resources; supply, doctor set, nurse set, technician set surgical equipment set, and operating room set. The last two have capacities of five. Time required for the process is based on a triangular distribution with a minimum of 112.6 minutes, a maximum of 118.9 and a most likely of 115.8 minutes (T 112.6, 115.8, 118.6). These times are based on Marjamaa, et.al. (2009), and Van Berkel and Blake (2007). There is another Find Supply process associated with surgery, requiring a nurse and a technician and using T .5, 5, 15 minutes. Following surgery, the patient is moved to a recovery room. Moving requires a nurse and a technician. It takes T 5, 15, 30 minutes of time. Recovery itself requires the resources of supply, nurse, and recovery bed (e.g., a recovery room). It takes T 15, 30, 60 minutes of time. This process is also associated with a Find Supply process requiring a nurse and T .5, 5, 20 minutes. There is a decision point here, since one of the main purposes of monitoring recovery is to identify signs of complications. A decision point has been inserted here that asks if there are complications. In 5% of the cases, complications arise from factors that could be prevented using RFID technology and are associated with improper procedures such as insufficient sterilization during surgery or incorrect dosages and drug procedures. This 5% is routed back into the system at the preoperative exam process. The remaining 95% enter the discharge process. This process requires the resource supplies and staff pack. The time required comes from DeBusk and Rangel (2004) and is a triangular distribution with a minimum of 56 minutes, a maximum of 112 and a most likely of 84 (T 56, 84, 112). Since supplies are required, there is also a Find Supply process using a nurse set and a staff set requiring T .5, 10, 45 minutes. The model completes with throughput times and other data collected in counting modules. In order to simulate the need to sterilize certain resources in between patients, the Failure module was used so that the resources were unavailable for a short period of time in between patients. Under Advanced Processes, Failures, two types of failures were created, short failures taking a normally distributed time with a mean of 3 minutes and standard deviation of 1 minute and a longer sterilization procedure requiring a mean of 15 minutes with a standard deviation of 5 minutes. Resources using the short sterilization time between patients were doctors, nurses, and technicians. Resources using the longer sterilization time were all the rooms and equipment packs. These processes are summarized in Table 2. A diagram of the model is shown in Figure 3. For clarification, there are four types of modules used in the model diagram shown in Figure 3. Basic flowchart symbols are used to illustrate various aspects of the model. The main modules are the process modules representing the tasks that require time and resources. These are represented by rectangular shapes. The diamond shapes represent decision points and determine the flow of entities through the model based on true/false statements. Assign modules and record modules are supporting modules that merely assign variable to entities and record the data collected. These stages neither require time nor consume any resources. Model With RFID The model including RFID assumes that the implementation of RFID will significantly improve the coordination and visibility of the processes required to outpatient surgical system. These benefits were seen in somewhat similar situations in the section on Related Work. To represent that coordination and visibility, the additional processes associated with Find Supply have been removed in the ‘with RFID’ model since all supplies will be tagged with RFID labels and their whereabouts constantly monitored in real time. There should be no additional time needed to locate them. The decision point located after the recovery process, complications, has been changed to indicate no complications due to post-operative infections created by improper procedures or insufficient sterilization. The remaining processes, times, and resources remain the same as in the base model. A diagram of this model is shown in Figure 4. [insert Figure 4 about here] Model Cost Assumptions The models incorporate cost assumptions gathered from public sources. Nurses, administrative staff, and technical staff are paid on an hourly basis. Each set of wages was taken from the United States Bureau of Labor Statistics website (www.bls.gov). Wages used were $34.00/hour, $15.50/hour and $22.00/hour respectively. Doctors’ and Surgeons’ costs were based on a per procedure basis (identified as ‘per use’ in the table). While some hospital physicians are salaried (e.g., the Mayo Clinic in Rochester, MN), these models assume a more traditional form using the per procedure basis. The amount was based on an average of the billed costs of 72 common procedures across the United States as published by the American Medical Association (AMA) in its report located on the U.S. Health and Human Services website (http://www.cms.hhs.gov/HealthCareConInit/04_Physician.asp). This average amount was calculated as $100 per procedure (‘per use’ in the table) for the preliminary examination by a doctor and $2,071 per operation procedure (‘per use’ in the table) for surgeons. The average total reimbursable cost of an outpatient hospital stay was determined by the United States Office of Management and Budget to be $6,777.50 per day according to a report it published on the U.S. Government White House web site (Whitehouse, U.S. 2011). Subtracting the physician charge from the total daily charge leaves $4,706.5 which is assumed to be the amount needed to cover overhead (room charges), equipment needs, and supplies. This amount was therefore equally divided between the following on a per procedure basis: surgical equipment pack, lab equipment pack, exam room, operating room, testing room, recovery room, and supplies. Each of these items was assigned a cost of $784.42 per procedure. While the equal allocations may not be accurate in a real-life setting, they should suffice until empirical data can be found or collected. This is due to the fact that since the costs are allocated on a per procedure basis and the number of procedures per day should not vary significantly between the two models, differences in the allocations will not significantly affect the results of the model. The costs associated with each resource are summarized in Table 3. [insert Table 3 about here] Other Model Assumptions It was assumed that, because of the nature of the process, and based on interviews with hospital personnel, resources needed to be cleaned and sterilized between patients. This was modeled as resource downtime failures. People, such as surgeons, physicians and nurses as well as equipment and rooms were all assumed to need sterilization between uses. People used a shortened sterilization period (a normal distribution with mean of 3 minutes and standard deviation of 1 minute) and rooms and equipment a longer period (a normal distribution with mean of 5 minutes and standard deviation of 5 minutes). One of the advantages seen with RFID is that o

******CLICK ORDER NOW BELOW AND OUR WRITERS WILL WRITE AN ANSWER TO THIS ASSIGNMENT OR ANY OTHER ASSIGNMENT, DISCUSSION, ESSAY, HOMEWORK OR QUESTION YOU MAY HAVE. OUR PAPERS ARE PLAGIARISM FREE*******."