Patient falls represent the most frequent reported incidents in our 541-bed urban public hospital, reaching more than 200 occurrences per year. This prompted a fall-prevention program, consisting of several steps: 1) Descriptive analysis of 295 consecutive falls in order to look at the factors commonly associated with falls, and the physical, psychic, and pathological characteristics of patients, medication, circumstances, and environmental hazards; 2) A case-control study of ten high-risk medicine and surgery wards (178 patients), designed to identify which factors are discriminant to predict the falls; 3) A proposal for a fall-risk assessment score to be calculated at the admission of the patient; 5) If the risk is confirmed, implementation of general and specific actions identified by the components of the score. The score is based on a 15-point scale including: age greater than 65 years; history of previous falls; weakness or insufficient weight; impaired mobility or altered feet state; psychic disorders (depression, agitation, or risky behavior); neuro-psychiatric diseases (CVA, confusion, or dementia); fever or infection; and polypharmacy. The mean scores of fallers and control patients were 7.53 ± 3.02 and 4.81 ± 2.93 respectively (p < 0.000001). A score between 5 and 11 was chosen to start the fall prevention program, which may predict a large proportion (about 80%) of valid patients prone to falls in the assessed medical and surgical wards (scores higher than 11 correspond to severely diseased, often bedridden invalid patients, with little risk of falling). However, these criteria are not suitable for nursing homes and for long-stay patients.