Evaluation of an Prison Information System Failure


Today’s prisons are driven by the availability of information . From initial intake to final release, virtually all key decisions are largely driven by the availability of quality data, and careful analysis of data to support the variety of decisions made by administrators and personnel operating the prison (Brennon , Dave & Carr 2003a ). It is not surprising therefore that majority of prison administrators rely on the availability of Information System (IS), whether manual or computerized to provide the critical information needed for making important decisions within the prison system.

Information developed in prisons can have an impact on past and future cases, family closure, preventing future crimes, and ultimately to save lives both in and out of prison( Matthews). Interestingly, however, in many jurisdictions these information systems fail with devastating consequences ( Kaste 2016). Heek (2002) argues that the failure of public Information Systems are particularly acute in developing countries such as St. Lucia, with a majority of information systems facing either total or partial failure.

Using a number of analytical models, this paper will explore the reasons for the partial failure of the Bordelais Information Management System.

Bordelais Correctional Facility is the lone prison in Saint Lucia (St. Lucia). The Prison is located on the east coast of the island, near the village of Dennery. Bordelais was built in 2003 to replace the historic, Her Majesty’s Prison or Her Majesty’s Gaol, was built in 1827 to accommodate 60 inmates. However, at the time of decommissioning in 2003, the facility housed 400 inmates in abysmal conditions ((Dee Lundy Charles, 2014). Bordelais is a department of the Ministry of Home Affairs of the Government of St. Lucia ( St. Lucia- Access Government 2019a). Designed to house 500 inmates, as of April 2019, the facility had a population of 507.

The mission of the facility is “To protect society by providing a controlled, secure, safe, humane, productive and rehabilitative environment for those assigned to its custody’ (St. Lucia- Access Government 2019b ). In executing this mission the facility rely on the availability of information to guide operations such as inmate classification, inmate movements, inmate health, descriptions of inmate property, operational procedures, head counts, incidents that occur while an individual is in custody, etc (Elias 2007a). In 2016 the department contracted the services of an outside firm to design and implement a computerized information system to replace its existing manual information system ( Caribsecure 2019 ). The author had been working at the selected organization from 2013 as System Administrator and served as project leader in the implementation of the system.

Bordelais Information Management System (BIMS) seeks to integrate prison data into a single integrated system, which will provide prison administrators with real-time information on all aspects of prison management including inmate personal property, biographic information, visitor activity, mental and medical health information, inmate movement in and out the facility, case and crime information, remission, classification, parole, and inmate rehabilitation (Manzar 2013). BIMS’ rollout took place in 2017, one year after work commenced on the system. BIMS is a distributed multi-tier web application built on the principles of a Decision Support System (DSS) (Sarngadharan & Minimol 2009).

As highlighted by Heeks and McCort (2019), information produced in an information system can be used for the purpose of helping to make decision. BIMS is designed with the intent of providing quality information to the administrators of the facility, which in turn assist in making decisions on the way to manage the institution. Using the Onion Ring Model (figure 1) as a descriptive model, we can see that information is at the core of the Bordelais Information Management System (Heek & McCourt, 2019). In the prison context, policymakers should obtain perspectives on past trends, present levels, and likely future trends of any practice or problem in the jail and local criminal justice system. It is also important to distinguish between causes, correlates, and consequences of any jail procedure, problem, or trend (Brennon , Dave & Carr 2013b ).

Data-driven decision making at the facility relies on quality on practices (e.g., quality control by qualified management, organizational learning, and continuous improvement) that have originated mainly in industry and are designed to support both decision-making and planning at all levels of the organization. The goal of all of these approaches is organizational improvement by the systematic collection and use of categories of data that broadly reflect the functions of the organization at large as well as specific departments (Brennon , Dave & Carr 2013d ).

In an effort to provide a more detailed description of BIMS we will use CIPSODA Model (figure 2) as an aid for describing the new system from (Heeks and McCort 2019).

Data capture

Brennon , Dave & Carr (2013) rightly highlight that It is critical to realize that the types of data captured in the system will vary according to the units or departments of the facility and for the individual functional unit under consideration (e.g., security, medical and programming, staff resources). The categories of data collected in the system includes:

  • Input data, such as equipment and costs of the labor, facility, and programs.
  • Outputs or work done, such as treatments provided, classifications completed, number of supervision tasks completed, number of admissions completed, and number of criminal history searches.
  • Results or outcomes data such as escape rates, disciplinary rates, rates of inmate injuries, and staff morale levels.
  • Work quality, such as error rates in data collection, percentage of tasks completed on time, numbers of inmates mis-housed in wrong custody levels, and rates of compliance with various correctional standards.


These details are entered into the system during the input stage as illustrated by the CIPSODAR model. The details entered into the system is captured in various paper forms and diaries that must be completed by law (Ilo.org, 2017) . These include intake form, medical form, movement diaries, Sentence order, etc.

Processing Data

In this step, raw data are turned into policy-relevant information or actionable knowledge (Brennon , Dave & Carr 2013e ). At this state the data collected is manipulated in a number of different ways, such as sorting, correlating, grouping, etc. Brennon , Dave & Carr (2013) suggest a critical issue at this stage is the analytical capacity and skills of the operators of the system, making sense of the raw data and transforming it into useful knowledge at this stage requires some technical skills and training.

Storage and Output

Both the raw data and the refined information are stored on database servers. The processed information is stored in the form of reports that is available to the different units at the facility, such as management, medical unit, intel unit, and other units at the facility.

Making Decisions

Decision makers must rely on actionable knowledge, combined with their judgment and their knowledge of the jail, to select actions to resolve a specific problem. These actions may include revising the jail’s goals, assessing inmates’ needs when planning programs, designing responses to issues of noncompliance, evaluating the effectiveness of various programs, reallocating resources (Brennon , Dave & Carr 2013f ).


Management review policies and make changes to policy documents, these policy documents are then passed on to the implementing units for implementation. This is often done after quarterly management meetings.


At the next management meeting unit heads, supervisors and other managers report on the implementation of the new policies. At these quarterly management it is decided whether the actions taken is producing the desired results. Due to the nature of the institution and the nature of the solution, it may be necessary to allow a longer period than one quarter to determine the impact of the desired solution.

At the core of the information system is a web application. The web application in BIMS was created using the ASP.Net web development platform provided by Microsoft. HTML is used for the visible web contents, javascript, ASP.Net for server scripting, a SQL database is used to store all data. The application is accessed by users through a Web browser. The browser creates HTTP request for specific URSs that map to resources on a Web server. The server renders and returns HTML pages to the client, which the browser can display. The core of the application is the server-side logic ((Meier et al., 2009)) which handles the intelligence of the system. Initially, the application will serve users on Bordelais’ intranet, and later move to the internet with the implementation of the Carib Secure project (Carib Secure 2019). The application was developed to require minimum hardware requirement.


BIS utilizes a multi-tier client/server platform, where the Web server is deployed into a perimeter network and accesses the application. A firewall is implemented between the client and the Web tier, and another firewall between the Web tier and the application or business logic tier (Weier et al 2009) as illustrated in figure 4. In this distributed deployment client devices (PC/Tablet/Phone) access web forms on the Dell PowerEdge T20 Server, two additional Dell PowerEdge T20 Server serve as application server and database server (Ernest 2015). This architecture was implemented for security, scalability, and load balancing (Weier et al 2009)

The measurement of information systems success or effectiveness is critical to our understanding of the value and efficacy of IS management actions and IS investments (DeLone & McLean 2002). There are a number of different approaches to assessing the success of information system. One of the most popular and validated measure is the DeLone and McLean (1992) IS success model (Ojo, 2016). The model, which was originally released in 1992 and subsequently updated in DeLone & McLean (2003) is shown in figure 5 below (Heek & McCort 2019).

Information Quality

Information quality seriously impacts on the efficiency and effectiveness of organizations and businesses, this particularly so in the field of low enforcement and corrections (Lindsay, Jackson and Cooke, 2010). Information quality plays an important role in the area of prisoner information, court related information, remission management, prisoner admission and release, detention of under-trials, prisoners work allotment and monitoring, prisoners medical management, victim compensation, parole, etc (Manzar 2013).

The failure to provide accurate, and timely data where these areas are concerned can result potentially in prisoners being released before their due date, the wrong prisoner being released, prisoners not undergoing court stipulated treatments, and so on. The system relies on officer to enter information in a timely manner. Due to staff shortage and indiscipline among officers data is often entered into the system sometimes days or weeks after an event, as a result the system does not always have the most accurate data. Additionally, supervisors that are responsible for verifying the data entered experience considerable difficulty in using the client computers provided (lack of proper training was highlighted the main reason for this, followed by an unwillingness to depend too much on computers).

System Quality

The system did not take into consideration that majority of the records at the facility must be kept on paper. As a result it creates some duplication of effort, which results in staff being reluctant to enter data into the system. Staff members also find recording information the ‘old way’ is faster, especially at the admission department where a large volume of inmates passes through on a daily basis. Although the system is relatively fast in accepting and processing data, users find it difficult to navigate to the various screens to input the required data. They complain that the new system is too different from the forms they are accustomed to. The accuracy of the system depends on the officer entering the initial data and the supervisor who must validate the data, often due to staff shortages both officer and supervisor neglect to populate the system. it is hoped that as users become more familiar with the system and recognizes the benefits they will be more accepting of the new system.

Service Quality

Delone and McLean (2003) put forward that poor user support, whether provided internally by the IS support department or externally, play a significant role in user acceptance of information systems. The support delivery for BIMS was initially provided by the contractor during the implementation period. During this period the contractor conducted formal training sessions where user concerns were addressed. However, at the end of the training the contractor introduced a system where all support calls were handle one day in the week. This meant that if the contractor came in on Mondays and the user experienced a problem, he/she would have to wait until the following week to receive support. This caused significant frustration among users of the system.

Intention To Use

Information systems cannot improve organizational performance if they aren’t used. Unfortunately, resistance to end-user systems by managers and professionals is a widespread problem (Davis, Bagozzi and Warshaw, 1989). In the case of the information system at Bordelais, although the system can work, end-users are demonstrating an unwilling to use the system. When questioned, many users highlight the duplication of work as the main reason for their resistance. The longer the system remain unused the more difficult it will be for users to use the new system.

Use And User Satisfaction

One of the fundamental information system success criteria is user satisfaction. Evaluating the satisfaction and its determinant is an important exercise to gauge the value and effectiveness of the information system investment (Kassim et al., 2012). There is a clear indication that users are not satisfied with the system and have not fully accepted the system (Wu & Chen, 2005). Initially, there were pocket of users who were enthusiastic about the introduction of the new system. However, with the multiplicity of issues encountered, these users are now using the system mostly as a result of pressure from upper management. Interestingly, majority of users see some value in the new system, but complain that they (users) were not widely engaged during the design and implementation of the system.

Net Benefits

Staff and inmate safety. A first and central role of a jail is to provide valid identification of dangerous offenders. This identification relies on carefully collected, individual inmate background and risk factor data (e.g., criminal history, past convictions, arrests, past behavior problems) coupled with appropriate classification and housing procedures. Both inmate and staff safety rely on valid identification, classification, separation, and supervision of dangerous inmates. Public safety.

A second role of the jail is to provide public safety. This requires effective classification, housing, supervision, and inmate management strategies that may reduce the risk of escapes, walkaways from work assignments, new crimes committed while on work release, recidivism, erroneous community placement, and so on. Protection against liability and protection of inmates’ rights. A third role of the jail is to minimize liability and avoid costly lawsuits and monetary awards. In addition to the provision of a safe environment, jails must provide a quality of life that ensures access to and meets the needs of inmates’ medical, dental, mental health, nutrition, and clothing needs (Brennan, Carr and Wells 2013) .

In all these areas accurate, readily available data that is easily and quickly accessible is vital. It is in the interest of the facility, nation, inmates, and officers that the issues confronting the system be resolved and fully implemented.