The purpose of the study was to investigate the clinical effect of a clinical pharmacist (CP) intervention upon admission to hospital on inpatient harm and to assess a potential educational bias. Over 16 months, 593 adult patients taking ≥4 medications daily were included from three Danish acute medicine wards. Patients were randomized to either the CP intervention or the usual care (prospective control). To assess a potential educational bias, a retrospective control group was formed by randomization. The CP intervention comprised medication history, medication reconciliation, medication review and entry of proposed prescriptions into the electronic prescribing system. The primary outcome of inpatient harm was identified using triggers from the Institute of Healthcare Improvement Global Trigger Tool. Harms were validated and rated for severity by two independent and blinded outcome panels. Secondary end-points were harms per patient, length of hospital stay, readmissions and 1-year mortality. Harm affected 11% of the patients in the intervention group compared to 17% in the combined control group, odds ratio (OR) 0.57 (CI 0.32–1.02, p = 0.06). The incidence of harm was similar in the intervention and prospective control groups, OR 0.80 (CI 0.40–1.59, p = 0.52) but occurred less frequently in the intervention than in the retrospective control group OR 0.46 (CI 0.25–0.85, p = 0.01). An educational bias from the intervention to the control group might have contributed to this negative outcome. In conclusion, the CP intervention at admission to hospital had no statistically significant effect on inpatient harm.