Organizations have a need to ensure that all its products / services produced free of defects to avoid costs of recovery, remanufacturing and / or waste and customer returns, as well as possible damage to the brand image for the company. However, inspect all products / services would be prohibitively expensive. SPC is possible to ensure that 100% of products / services are produced without defects without control 100%.
SPC is not applied to control products manufactured, but for preventing the occurrence of defects during production.
Actually sampling requiring SPC is not expensive but is a preventive control so that no defective products are manufactured. SPC is a tool read it provides a scientific basis for defining optimum controls replacing the need to inspect all products, so that control costs while ensuring provided are minimized to almost 100% of the delivered products are free of defects.
For tangible products, if a manufacturer does not apply tools such as SPC and Poke-Yokes, defects will occur, but will have a second chance control 100% of the manufactured products to detect and prevent delivery of defects customer; at least the image of the company would be saved. However, in the case of utilities, there is no second chance because the services are delivered as they occur. If the defect in service is hazardous to the user, prevention is the only option and SPC can help achieve this without spending 100% control during production of the service. For example, in a radiotherapy treatment, if patients are not well placed in the machine, it would radiate and destroy healthy tissues instead of destroying tumor tissues.
The problem of positioning error in radiotherapy treatment is comparable to the problem in the industrial sector, since it must reproduce patient positioning in each of the multiple sessions to protect healthy tissue and ensure the effectiveness of treatment. But verification of placement in each treatment session not only prohibitively expensive but also generate a period longer waiting for new patients in the public health system in countries like Spain, where there is always more demand than offer.
Following this conflict between quality and productivity in most radiotherapy centers verification 1 in 5 treatment sessions is performed. This type of sampling has a very limited benefit because it can only be corrected the day when the error is measured. With the SPC method, if it could be demonstrated by appropriate statistical sample, the process was stable throughout the shift of treatment, one could say with 99.73% probability that all patients treated during the turn were placed correctly .
In any process variability occurs. Errors due to common causes always present in any process. First, the variability is quantified due to common causes to see if it is acceptable, if not, an improved process is done to reduce this variability. Having found the variability of the process generating an acceptable range, you can control the stability of the process, which, when the process is stable, can be said that the process has varied only within the acceptable range.
In order to control the stability of the process control limits [: Average X ± 3 σ average process] are used. If the process loses stability will be due to excessive variation produced by a «assignable cause» (or «special case»); therefore, it identifies the assignable cause, to take corrective action and thus redirect the process towards its original variability.
In a radio therapy center treating patients with prostate cancer SPC methodology was used to define the control protocol patient placement. the positioning error was measured three times during each session (at 8, 10 and 12 hours) in three groups of three patients each. Each group consisted of three consecutive patients. The result of the Media process values (process average) of each subgroup was drawn in the control chart (see Figure). The chart range (R) is not traced to simpler control in real time (to cause minimal disruption during treatment, and avoid delays or loss of productivity), but technicians were also instructed to control Range ( plot on a graph); they were told not to treat any patient whose positioning error went beyond 5 mm, and asked that in such cases, call the physicist who provide them assistance. Thus, technicians calculated the average subgroup only when the individual values were not particularly high (<5 mm). Thus, the range of variability was kept under control. Technicians were trained to perform control charts with X.
The existing weekly monitoring in many radiotherapy centers is currently based on the reactive philosophy and not enough. Statistical Process Control is based on the understanding of the variability in ensuring that the variability is acceptable (if not, should reduce variability by improvement actions), and finally control the stability of the process, always within an acceptable range. This is a proactive and preventive strategy of quality control.
The existing current method is based on weekly sampling control without statistical bases. If considered a weekly control in a patient 26 treatment sessions, one can ensure that in 4 of the 26 sessions placement was correct. In other words, you can only ensure that 15% of the sessions of each of the patients are free of positioning errors. You do not have any idea how many placement errors occurred during the remaining 22 sessions. Does this mean no more risk of mortality and morbidity with less chance of surviving cancer? Ethically it is complicated to allow this to continue, but as you can not make controls 100% of the sessions, is being allowed.
Another problem that currently exist in many radiotherapy centers not know exactly what level placement error (due to the common cause of variability). Each radiation therapist uses their own judgment to decide what level of error detected placement is acceptable or needs correction. Statistical Process Control is a good statistical tool that can help understand and control the process variability and prevent it varies beyond the acceptable range.
The challenge is to find statistical basis for preventing errors without increasing control costs. Statistical Process Control provides an alternative method that seeks to achieve proper placement without increasing the number of controls. With the existing method, if 4 total patient controls along its treatment are made, assuming that treatment is provided to 100 patients per month, a total of 400 controls. With the proposal of Statistical Process Control if 3 subgroups control are performed per shift, with each group consisting of 3 consecutive patients, 396 checks per month are performed. Thus, with the same total number of controls, the SPC method gives the guarantee that all patients during all sessions are free placement error while the current control ensures only 15% of the sessions!
In fact, more savings is possible once better understanding of how varied the process and what factors cause the process to lose stability. If frequent cause of loss of stability is well understood, it could get to consider reducing the number of controls from 3 turn 2 can lower total controls from 396 to 264. There are still more potential to reduce the number controls for shift 396-176 if subgroup size 3 is reduced to two (the minimum allowed) and thus save even more controls, without sacrificing the probability of 99.73% of all patients placed in a process stable are within acceptable error. With the funding problems that have public health, SPC allows an intelligent quality control that eliminates the costs and risks at a time.
Rajaram Govindarajan is Six Sigma Black Belt and Chief Auditor SPG ( www.certificadoISO9001.com ) and is implemented in SPG policy «add value» through audits ISO9001 so that their customers generate a culture of quality. For the full study, see Rajaram Govindarajan et al., Statistical process control can help prevent treatment errors without increasing costs radiotherapy, Rev Care Quality. 2010; 25: 281-90.