From Logging to Learning: How PD Care System Helps Builds Self-Management Skills One Day at a Time

Patients log activities across seven categories that make up most of a day: sleep, work, housework, leisure, exercise, activities of daily living, and rest. These were chosen because they cover the domains where pain and behavior interact in daily life, whether patients think about it or not.

For each activity, patients record three things, duration, intensity level and how rewarding it was, e.g. if it gave value to the day. That last one matters most. As we discussed in our post on ChatGPT Health and specialized digital therapeutics, wearables and general AI tools work with physiological signals, heart rate, steps, sleep stages, but miss the context needed for useful pain management guidance. PD Care System is designed to fill that gap.

A wearable measure 45 minutes of elevated heart rate. In PD Care System the patient logs 45 minutes of gardening, which was moderately hard, but really enjoyable, and later the same evening, an evaluation of how much pain interfered with the day. One tells you what the body did. The other tells you how daily life connects to pain. Only the second can guide the user to meaningful, individualized recommendations

Patients also note specific tasks within each category, washing dishes, walking the dog, a work meeting, playing with their kids, building the kind of detail that makes personal patterns visible and gives the AI engine something genuinely useful to work with.

After about a week of logging, the PD Care app begins generating individualized, activity-based insights drawn from each patient's own patterns. These are not generic recommendations like “exercise more”, or standard pacing guidelines. Instead, they are structured, data-driven observations, contextualized into daily recommendations made from the previous days' reported activities including duration, intensity, satisfaction, and pain level, analyzed across multiple time windows.

The system also runs a machine-learning model to estimate pain interference, i.e., how much pain may be affecting daily functioning, based on the patient's longitudinal inputs such as activity exposure, pain ratings, sleep satisfaction, and related self-reports. This adds a functional perspective alongside pain intensity. Users receive insights both as text and as graphs in the My journey view. Together, these visuals and observations can help patients distinguish between patterns such as "high pain with low disruption" on some days versus "moderate pain with high interference" on others. Recognizing these nuances can support more focused and productive pacing discussions during clinical visits.

Examples of the types of insights patients receive include patterns where periods of higher work intensity tend to be followed by higher pain, moderate leisure intensity linking to lower same-day pain, or better sleep satisfaction connecting to lower pain later in the week. These patterns are not instructions or rules—they are prompts for reflection, helping individuals notice how different aspects of daily life relate to pain and supporting more flexible, values-aligned choices.

The clinical pain literature is consistent on this point, effective pacing depends on individual thresholds, not universal activity rules. Two patients with similar diagnoses can have quite different tolerances for the same activity load, and within the same patient, tolerance shifts with sleep quality, stress, cumulative load, and day-to-day demands. Making personalization work in practice requires structured, longitudinal data that captures both exposure (what the patient does) and response (how pain and function are affected).

The "Shape Today" function in the app translates these individualized associations into a brief reflection exercise in the form of a daily plan. It offers daily planning suggestions based on observed patterns and the estimated pain-interference signal, highlighting where adjusting duration or intensity might help support a more sustainable day. The tool is designed to support shared decision-making, not replace it: the patient stays in control, weighing insights against their own plans and priorities.

Over time, PD Care System supports a structured learning cycle: monitor activity, identify patterns, test adjustments, track outcomes, and refine the patient's understanding of their own thresholds and when they can confidently be more active. This iterative approach aligns with the mechanisms that research links to improvements, including strengthened self-efficacy and improved self-regulation of activity pacing.

For clinicians, the PD Care System extends insight into the period between visits. Through the Remote Therapeutic Monitoring (RTM) dashboard, providers can review longitudinal patterns of activity and engagement. This can help identify potential early signs of activity avoidance or reduced adherence, observe changes in sleep patterns or pain reports, and consider objective behavioral data alongside traditional patient-reported outcome measures.

This approach may support more continuous, longitudinal management of chronic pain and provide a stronger foundation for individualized, collaborative treatment discussions.

Another feature of the system reflects its foundation in Acceptance and Commitment Therapy (ACT). On days when patients report low energy, the app encourages a brief pause for reflection, not to push through discomfort automatically, nor to withdraw prematurely, but to check in with their current capacity and make choices aligned with personally meaningful values.

At times, this may involve engaging in an activity despite fatigue because it supports social connection or other valued life domains. At other times, it may involve prioritizing rest in order to sustain functioning over the longer term. This reflects psychological flexibility in practice, the capacity to make intentional, value-guided decisions while acknowledging the presence of pain.

The system also includes relaxation and mindfulness-based exercises intended to help regulate stress responses that can amplify pain perception, as well as a quick-access feature designed to support coping during pain flare-ups.

Throughout, the patient remains the active decision-maker. The system is designed to provide structure, data-informed feedback, and guidance, recognizing that algorithmic recommendations are dependent on the quality and consistency of the data provided.