BACKGROUND: Early prediction of therapeutic response can optimize treatment strategies in atopic dermatitis (AD). Baricitinib is approved for moderate-to-severe AD in Europe, Japan and other countries. OBJECTIVES: To identify early clinical improvements that can reliably predict a later clinical response to baricitinib in adults with moderate-to-severe AD. METHODS: Using data from one topical corticosteroid combination study [BREEZE-AD7 (NCT03733301)] and data pooled from two monotherapy studies [(BREEZE-AD1 (NCT03334396) and BREEZE-AD2 (NCT03334422)], we calculated the sensitivity and specificity, along with the positive predictive value (PPV) and negative predictive value (NPV), of predefined changes in single and combined clinical scores at weeks 2, 4 and 8, to predict clinical response at week 16. Clinical response was defined as ≥ 75% improvement in Eczema Area and Severity Index (EASI 75), ≥ 4-point improvement in Itch Numeric Rating Scale (Itch NRS ≥ 4), or a combination of both. RESULTS: Composite predictors had higher predictive accuracy for week 16 response outcomes than did single parameters. This was evident as early as week 4 for the combination of EASI 50 or Itch NRS ≥ 3 and of validated Investigator Global Assessment for AD (vIGA-AD) score ≤ 2 or Itch NRS ≥ 3 (sensitivity 87-100%; NPV 68-100%). The predictive accuracy of these composite clinical predictors for week 16 response outcomes was highest at week 8 (sensitivity 92-100%; NPV 80-100%). At both weeks 4 and 8, EASI 50 or Itch NRS ≥ 3 had higher sensitivity and NPV than did vIGA-AD score ≤ 2 or Itch NRS ≥ 3. CONCLUSIONS: Improvement in signs and symptoms early during treatment with baricitinib 4 mg once daily predicts clinical response at week 16, providing a tool for dermatologists when choosing treatment strategies for patients with moderate-to-severe AD.