Skip to content Skip to navigation

Intelligent software: A roadmap to recovery

November 1, 2007
by Pieter Hubbard
| Reprints

When software first came to the addiction treatment community in the 1980s, it was little more than an electronic version of paper forms. While the software may have been smart enough not to ask the same question multiple times, it had very little intelligence built into it. With the addition of accreditation to clinical processes in the 1990s, the requirements to know more about the people we serve grew tremendously. Electronic forms were replaced by databases collecting information about patient performance, retention, and outcomes. The software made many advances but still was viewed as a tool, with the teacher being the end user. Today there exists a new generation of software, intelligent software, and it has dramatically changed those roles.

Improved design

One of the hardest things to share with people new to the treatment community is the clinical knowledge one acquires from years of experience. With so many programs facing budget cutbacks, many facilities don't have the resources or time to train their staff. This is where the latest generation of software can help out. Using the newest technology available, software can react to a user, changing prompts and options based on the answers the user provides.

The software can look at a particular prompt, and if the response indicates the need for more in-depth information, additional questions will systematically present. For example, if during an assessment interview a question posed to the patient is, “How often do you feel sad?” and the patient responds, “Rarely,” the clinician might not want to pursue the line of questioning much further. If the patient responds, “Often,” the clinician might want to follow up with more questions about depression. However, if the clinician is new to the industry, how would he/she know which line of questioning is appropriate or which additional assessment tools are appropriate to use? How should the questions or tools be introduced to the interview process to make them appear like a natural continuation of the clinician's conversation with the patient?

Well-designed software will look at the responses selected by the user, and based upon a decision tree skip ahead to a new area of questioning or display additional questions about the issue. In the above example, if the response indicates additional questions are needed about depression, the software will display several follow-up questions. If the responses to those questions indicate that a depression issue exists, the software will prompt the user to use a specialized tool such as the Beck Depression Inventory to get a better understanding of the patient's level of depression. If the user agrees, the software will introduce the questions from the Beck tool into the assessment process. The outcome of the depression survey will be embedded into the assessment as the supporting documentation for concluding that the patient has an urgent need for depression treatment.

Benefits for treatment

There are three main advantages to having this type of artificial intelligence in your software system. First and foremost is having two sets of “eyes” on the patient, looking for indicators of treatment needs. A well-designed system will help the clinician accurately assess a patient's strengths and needs. The software will introduce evidence-based tools and best-practice standards at the appropriate time during the recovery process. The software will ensure that the issues uncovered by various assessment tools are included in the resulting treatment plan. The issues covered in the treatment plan will guide the clinician through the appropriate topics to discuss in individual and group sessions.

This is not to suggest that the software will “think” for the clinician. It is a tool, designed to remind the clinician where on the road to recovery the patient is at any point in time.

A second benefit of having intelligent software is the software's ability to follow a clinic's policies and procedures for treatment, rather than the clinic following the software vendor's requirements for data entry. Because intelligent software uses decision-making trees to process information, clinic staff are involved in a configuration process to set up which tools the software uses, when are they used, and what responses trigger additional processes. The end result is that the clinic's treatment protocols, policies, and procedures become embedded into the software.

Finally, intelligent software becomes a teaching tool. An inexperienced clinician might not recognize a pattern of responses as an indicator of a bigger need, a trigger for relapse, or a potentially life-threatening situation. The software will teach the clinician to think like a seasoned veteran. It will guide the user to ask appropriate questions and introduce additional sources of information that the clinician might not have been aware of. The advantage to the clinician is the exposure to best-practice standards and evidence-based treatment protocols. The software also teaches the user about the importance of linking information discovered through the assessment process to the ensuing treatment plans and session notes. As a result, the treatment process is more personalized to suit each patient's individual needs.

The clear winner is the patient. Clinicians might begin to use a wider variety of tools to chart a course for recovery. Less-experienced clinicians will become better at accurately assessing a patient's needs and strengths. And as the recovery process is tailored to those needs, patient retention in treatment and patient outcomes will only improve.

Pieter Hubbard is the Manager of Methadone Services for Tower Systems, Inc., a California software development company whose products include a software package tailored to management of methadone clinics. His e-mail address is

Pages

Topics