FAQs: What are ADSMs and how are they used in Limbic?

Kicking off Limbic's FAQ series, I'm here to guide you through some of the most pressing queries you might have about how our products function. Picture this as Limbic's version of an open office hour, where no question is too big or small.

From the difference between wellness and clinical healthcare apps to breaking down clinical best practices and understanding how Limbic leverages Large Language Models (LLMs). We will help distill and clarify your burning questions about Limbic because AI should be explainable and accessible for everyone. 

This post discusses the critical role of Anxiety Disorder Specific Measures (ADSMs) in enhancing problem identification and treatment outcomes within NHS Talking Therapies services. I will also explain how we use ADSMs in Limbic Access.

What are Anxiety Disorder Specific Measures (ADSM) questionnaires?

ADSMs are questionnaires designed to identify the symptoms of specific anxiety disorders. Using these measures ensures that patients are placed on the correct treatment pathway for their anxiety, and their progress can be effectively evaluated by NHS Talking Therapies services. For instance, a specific Obsessive Compulsive Disorder (OCD) questionnaire is used to diagnose OCD.

A brief history of current practices 

• NHS Talking Therapies services employ a session-by-session outcome monitoring system using the PHQ-9 (Patient Healthcare Questionnaire) for depression and the GAD-7 (General Anxiety Disorder) for generalised anxiety. However, due to its general nature, the GAD-7 lacks items specific to disorders like social phobia, agoraphobia, OCD, panic disorder, and PTSD (Post Traumatic Stress Disorder) (Clark, 2017), and so its predictive power is poor.

• Since 2010, national guidance has recommended using well-validated ADSMs alongside the PHQ-9 and GAD-7 to address this gap (Clark, 2017).

Imagine each anxiety disorder as a door, locked with its own unique key. ADSMs are similar to a meticulously crafted set of keys, each designed to unlock the door to effective treatment and recovery for a specific type of anxiety disorder.

Impact of not using ADSMs

Despite recommendations, ADSM use is low in NHS Talking Therapies. In 2017, David Clark published a report on the underuse of ADSMs, suggesting that this can lead to premature discharge and overly optimistic recovery rates. The report specifically highlights discrepancies in recovery rates when using general measures versus ADSMs for social phobia and PTSD, indicating that specific measures are crucial for accurate assessment and treatment planning (Clark, 2017).

Why have we included ADSMs in Limbic?

National guidance has recommended that ADSMs are completed at assessment; however, ADSMs are used in less than 20% of eligible cases (Clark, 2017). Not using ADSMs can have a significant impact on patients, for example, with an incorrect diagnosis or early discharge.

One of the potential reasons for low ADSM use is that clinicians simply don’t have the time to implement them, as assessments are already packed with information collection. To support clinicians, Limbic can collect the following six ADSMs at referral:

• Social anxiety: Social Phobia Inventory (SPIN)

• Panic disorder: Panic Disorder Severity Scale (PDSS)

• OCD: Obsessive-Compulsive Inventory Revised (OCI-R)

• PTSD: PTSD Checklist for DSM-5 (PCL-5)

• Health anxiety: Health Anxiety Inventor (HAI-18)

• Specific phobia: Severity Measure for Specific Phobia—Adult

All of these measures are evidence-based and are included in the NHS Talking Therapies Manual. But administering six questionnaires for every patient is not feasible or time efficient, so how can we choose which ones to administer?

In Limbic Access ADSMs are chosen by a clever algorithm 

Our artificial intelligence algorithm is a probabilistic machine learning model that determines which ADSMs are asked. A probabilistic machine learning model is a type of model that incorporates probability theory to handle uncertainties in predictions. In other words, it can express uncertainty about its predictions, making it extremely useful in real-world scenarios where data is uncertain or incomplete. This algorithm has been rigorously trained and tested on data from >46,000 patients.  

This algorithm selects which ADSMs to administer based on a service user’s previous responses to:

• The goals for therapy and main problems they want to address 

• Minimum Dataset outcome measure questionnaire scores

• Behaviour, such as screen time and typing speed 

Based on this data, a maximum of two ADSMs are administered. In testing, our model has shown a 93.8% accuracy in selecting the correct ADSM across different diagnoses.

This cuts down the number of conditions that need to be excluded within the assessment, saving time and allowing for a higher quality of assessment with targeted questioning. 

Now let’s break it down with an example: 

1. A patient enters their problem description in their own words: ‘I have anxiety and panic attacks when I am in public spaces and when I have to talk to people I don’t know very well. Some examples include work meetings, weddings, and going places I don’t know.’

2. With basic information covered, Limbic Access gathers initial outcome measures as part of the Talking Therapies dataset. This includes: PHQ-9, GAD-7, Phobia Scale, and WSAS (Work and Social Adjustment Scale). With free text inputs and minimum data set outcome (MDS) measure scores, our artificial intelligence algorithm can now select the appropriate ADSMs. This form of tightly controlled, automated decision making is why Limbic Access is regulated as a Class IIa Medical device. Read more about this here. 

3. In this example, based on the MDS outcome measure data and the free text input, Limbic Access considers that this patient has anxiety, and therefore the probabilistic model chooses the following ADSMs to gather at the point of referral: PDSS & SPIN.

4. A summary of this information is uploaded into your Patient Management System.

As we aim to refine and improve mental health services, the use of ADSMs are an integral step towards more personalised and effective care. By embracing these measures, clinicians can offer treatments tailored to the individual, paving the way for more accurate assessments and better outcomes.

If you are curious to learn more about how Limbic Access improves the quality of assessments and treatment outcomes, explore the evidence here. 

-

David Clark (2017). ADSMs Recovery in IAPT. PC-MIS. Retrieved from https://www.pcmis.com/wp-content/uploads/2022/04/ADSMs_Recovery_IAPT.pdf