Posted at 10.13.2018
The computer aided despatch system was released on Oct 26th 1992 with the purpose of improving efficiency, controlling resources and reducing the employees requirements of the existing London Ambulance System (citation). On Oct 26th 1992, same day it was created, the CAD system failed.
Many literature have come out with reasons how the system failed, highlights which are (Sommerville, 2004):
Emergency response had taken a long time,
Communications with ambulances failed and therefore they were lost from the system
The system could not cope with the load placed onto it during normal use
Serious errors diagnosed in the procurement, design, execution and advantages of the machine.
To fully understand the CAD system inability, one would choose to give a brief history on the London Ambulance Service.
London Ambulance Service
The LAS was monitored by the South West Thames Regional Health Expert but became a Country wide Health Service (NHS) Trust on 1st of Apr, 1996 (LAS website 2011). It is the major Ambulance service in the world covering an area of over 600 rectangular miles and a human population of over 6. 8 million people. Typically, the LAS bears over 5000 patients everyday getting between two thousand and two thousand 500 calls each day out of which about 1500 are crisis telephone calls (Sommerville 2004).
A typical CAD system is shown below:
http://www. cs. ucl. ac. uk/staff/A. Finkelstein/advmsc/3. pdf
It mainly contains a CAD system, a marketing communications software, Radio system and mapping software amidst others and is used for taking cell phone calls, id, allocation and mobilization of resources.
This CAD system was released to displace the manual despatching system recently utilized by the LAS. Top features of the Manual System included (Sommerville 2004):
Call Taking - Telephone calls were documented on forms and delivered to central collection items on conveyor belts. Locations identified on maps.
Resource Id - Once the form is accumulated, it is approved onto a region based allocator who chooses on what resources have to be allocated, fills this in an application and is handed over to a dispatcher.
Resource Mobilization - this dispatcher communicates with the ambulance station who then delivers out the ambulance.
This entire process was expected to take less than three minutes.
http://www. cs. ucl. ac. uk/staff/A. Finkelstein/advmsc/3. pdf
The new system was likely to be fully computerized using resources more efficiently. The features include:
A CAD software
A mapping system
MDT terminals in ambulances
Automatic Vehicle Finding System (AVLS)
The specifications stage of a job is the level of which all stakeholders involved in the project get together to select certain key issues a few of which are:
Describe purpose, goals and deliverables.
State parameters (timescales, budgets, range, scope, place, authority).
State people engaged and what sort of team will continue to work (consistency of meetings decision-making process).
Establish intermittent break-points where progress is assessed and results measured.
It is clear that the LAS CAD project weren't explicit enough in the areas highlighted above. For example, some of the deliverables for the project were over ambitious. Despite the fact that existing CAD systems were looked into, the idea of adapting them was easily discarded as it had not been cost effective considering the differences in size of locations. However, no point out was manufactured from investigation of other CAD systems used by ambulance services in large cities all over the world.
In conditions of deadlines, a written report by Arthur Andersen suggested at least a 19 month time series for the delivery. This survey was also discarded and an eleven month deadline was establish.
Some other issues include non inclusion of stakeholders in the introduction of the specs, non addition of intermittent items in enough time for project review.
The System Requirements Specification (SRS) was completed in February 1991. It given the next:
The new system would include computer aided dispatch, computer map screen, and automated vehicle location.
Using CAD software, the completed system would require integration with several other systems, including a computerized radio communication system, an computerized vehicle location system, and a mobile data terminal system with the CAD software.
Though SRS called for integration with CAD software, it didn't appropriately specify in details, how the integration would be integrated.
The detailed check out the specification also confirmed that the new LAS CAD system would bring about a total change in the standard responsibilities of staffers in the LAS i. e. call centre and ambulance staff. Unfortunately, these models of staffers were not a part of the process of growing the SRS.
Apart from the 'flaws' highlighted above, the SRS produced by the job owners and the System Design Specs (SDC) assessed by the job managers had not been totally wrong. For specifications, in spite of the 'ambitious' deadline, the job was still considered as doable by 5 of the 17 companies that submitted bids for the job.
One substitute solution in the specification process that shines is that the LAS may have given room for the adaptability of other similar CAD systems utilized by ambulance services in virtually any large and developed city about the world. Adapting a similar technology would have kept time, money and perhaps wouldn't normally have concluded in inability.
Another alternative solution in the standards process is breaking the entire project down into different and smaller minuscule projects making it easier to control.
The major problem in conditions of Specs with the LASCAD job is at assigning the job to a inexperienced and small software company SO. Even though all other requirements of the specification stage were met, they were met in equipment. For instance, these were break tips within the project timeline to measure success or failure, we were holding not enforced as SO maintained postponing these deadlines which alone should have sounded an alarm bell to the project owners.
In all, the pressure to deliver the CAD software within a particular time frame affected the quality and contributed to the failure of the project. Incapability to totally enforce the requirements of the SRS and SDC also contributed to the inability of the task.
A project planning level usually involves the following stages - Job Timescales, the task team, the job critical path examination, Gantt charts, job financial planning and reporting and job contingency planning.
The contract for the LAS CAD system was given to a consortium created by Apricot Pcs, Datatrak, and Systems Options (SO). This consortium acquired the lowest bet and one of the five companies (out of 17 that submitted a bet) who said they could meet the 11 month deadline. Out of the 11 weeks, the consortium spent 5 calendar months in developing a comprehensive SRC departing only about six months for the job itself. The small amount of time for job development left virtually no time for project screening which could have open the likely problems the system would eventually face acquired it vanished on.
The consortium decided to take up PRINCE, a technique that they had no real experience in and even though there was training effective use cannot be achieved. In addition, as expected in PRINCE, the task owners, LAS, didn't assign any person in their staff full time to the project team. Also, an important feature of PRINCE was also not totally put in place - the job didn't have a complete time Quality Assurance team and only 1 part time employee performed QA for the task.
In terms of the financials, the consortium bid 937, 463 that was about 44% less than the next lowest bid which can probably be ascribed to the actual fact that Systems Options (SO), which was growing the CAD software, was a little software consulting company with no past experience in CAD systems development.
As is seen above, there were major problems with not fully implementing the task as a PRINCE project. Using a in your free time, non independent staff for QA posed a danger to the success of the job.
The question also arises if the right task management strategy was chosen. PRINCE as a strategy is as yet not known to work when there is a non adaptable deadline as in cases like this. A methodology more tuned to this is SCRUM.
Perhaps another problem lies in the task contingency planning area. In early stages in January 1992 when the job was to be shipped, SO performed a partial trials of the system, partial because the full system was not analyzed. Now these assessments did not take into consideration possible system mistakes and failures as should have been done and thus cannot make contingency programs should in the event a fault happened in true to life. An independent QA team could have properly verified and tested the system if there have been one.
In the Project planning stage, another solution could have been the use of a far more suited strategy - SCRUM, would have taken into account the time and financial constraints. SCRUM breaks the task down into different SPRINTS with every sprint centered on obtaining a preset goal. Product owners aren't permitted to change the task requirement during a sprint and the aim is to deliver the final product in just a specified period of time. Using SCRUM could have delivered the task as at when scheduled.
It is clear that the PRINCE technique was not implemented in full. This might have been because of the inexperience of the Project team in using PRINCE and also the need to save costs. A full time, 3rd party QA team must have been attached to the project to ensure enough testing and verification of the software.
Overall, one of the major reasons the LAS CAD job failed was because it was not executed as a task management task. Despite the fact that the managers followed PRINCE as a methodology, their inexperience in the use of PRINCE was glaring plus some of the essential top features of PRINCE as highlighted in parts above were either not totally implemented or totally inexistent. It has been said in a few literature that if the duty had been completely integrated as a task management job by experienced PRINCE users, even with the short timeframe and budget, it may have still been successful.
In final result, there a wide range of lessons to be learned from the failing of the LAS task. These lessons can be looked at from different phases mixed up in project. It really is clear that proper investigation into what the task will entail had not been carried out or not utilized by the task owners - no surprise the job was designated to the lowest bidders without concern on the feasibility of the project being executed at this cost.
Also the job owners didn't thoroughly monitor the project all through its development lifespan. Breakpoints in the project planning stage that could have exposed flaws were not properly utilised. That is an area for future project owners to study from.
For the project managers, one of these failures is based on the fact that they were a tiny software company and therefore had little if any experience in managing such big jobs. The project itself had not been organized as a job management project and what is more, the correct methodology was not used. This is evidently down to inexperience.
In addition, the glaring flaw is in the region of lack of expertise. SO evidently did not have knowledge in the job management way they adopted and although they trained their employees in this field, implementation inexperience damaged the planning and delivery eventually.
If these are viewed and corrected, it'll save LAS from future failures.