BACKGROUND: Fampridine is an effective treatment to improve ambulation for some multiple sclerosis (MS) patients. Remarkable discrepancies exist between responder rates in clinical trials and the proportion of patients continuing treatment in clinical practice. This may be related to clinical phenotypes of MS patients, and the influence of patient reported outcome (PRO) on treatment decision making. OBJECTIVE: To analyse responder rates to fampridine on ambulation and upper extremity function (UEF) and the influence on treatment decision making in different clinical subgroups in a real-world setting. METHODS: MS patients with ambulatory impairment treated with fampridine were included. Patients were subdivided based on disease duration, clinical phenotype, Expanded Disability Status Scale (EDSS), baseline walking speed, and presence of UEF impairment. Ambulatory response was assessed with the Timed 25-Foot Walk (T25FW, responder defined as ≥20% improvement) and with the MS Walking Scale (MSWS, responder defined as ≥8 points improvement) as a PRO. For patients also reporting impaired UEF, the Arm Function in MS Questionnaire (AMSQ, responder defined as ≥15 improvement) was the PRO of choice. Decision on treatment continuation was based on improvement of T25FW, MSWS and the clinicians' overall impression for improvement. RESULTS: In total 344 patients were included of which 75.3% continued treatment. More patients with a relapsing clinical phenotype continued treatment vs patients with a progressive phenotype (83.6 vs 68.6%, p < 0.01). A positive linear trend was found between severity of walking disability, as determined by baseline walking speed, and T25FW response (p < 0.01), while there was an inverse linear association between walking disability and MSWS response (p = 0.03). However, the proportion of patients continuing treatment was similar between subgroups of baseline walking speed. Impaired UEF was reported by 183 (66.5%) patients, of which 64 (39.3%) were AMSQ responders. Patients responding on AMSQ compared to non-responders, were also more frequently MSWS responders (82.8 vs 65.3%, p = 0.02), while response on T25FW was similar, and continued treatment more often (85.9 vs 70.7%, p = 0.04). This suggests an influence of PRO on treatment decision making. CONCLUSION: Responder rates and treatment continuation of fampridine differed between clinical subgroups of MS. PROs influenced treatment decision making of fampridine in clinical practice, particularly in patients with mild ambulatory impairment or those reporting UEF impairment. To some extent, these findings explain discrepancies found between clinical trials and clinical practice, and support the importance of subgroup analyses and incorporation of PROs in clinical trials. For clinical practice, using PROs to assess patients experience in conjunction with performance measures helps in treatment decision making.