👉 Does the amount of dopaminergic medication (LEDD) change how people with Parkinson’s walk, and does it reduce their risk of falling?

To explore this, we combined two approaches:

1. Causal analysis (IPTW weighting):
   This method balances groups of patients with different medication doses (low, medium, high), making them comparable as if they were in a randomized study.
   It allows us to estimate whether LEDD itself changes gait or fall risk, rather than differences being driven by age, disease severity, or other factors.

2. Machine Learning prediction:
   We trained models to predict whether a patient was likely to fall, based on both clinical and biomechanical data.
   These models also highlight which features (e.g., medication dose, motor severity, age) matter most for fall risk.

---

🔑 Key Findings

Gait

We found no consistent evidence that higher LEDD improves biomechanical walking parameters (such as gait speed, step length, stability indices).
Even after adjusting for confounders, differences remained small or non-significant.

Falls

A strong signal emerged for falls:

• Patients with high LEDD doses had a 5–10 times higher risk of falling compared to those on low doses.
• Patients on medium LEDD doses showed a possible increase in risk (~2x), but results were less certain.
• These findings were robust across multiple sensitivity checks, suggesting the effect is real and not an artifact.

Machine Learning models

• The best predictive model was XGBoost (AUC ~0.82).
• The most important predictors of falling were:
  • LEDD (dopaminergic dose)
  • Motor severity (UPDRS-III, Hoehn & Yahr stage)
  • Age
  • Body Mass Index (BMI)
  • Gait speed
• Taken together, medication dose and clinical severity drive most of the fall risk signal.

---

🩺 What does this mean for clinical practice?

Dopaminergic therapy is essential for controlling Parkinson’s symptoms.
However, higher medication doses do not necessarily improve walking biomechanics.

More importantly, high LEDD seems to increase the risk of falls, especially in younger or less impaired patients.
Clinicians should therefore balance the benefits of stronger symptom control against the elevated fall risk when titrating dopaminergic therapy.