Polypharmacy is an important issue in older patients affected by dementia because they are very vulnerable to the side effects of drugs’. Between October 2021 and September 2022, we randomly assessed 205 old-aged outpatients. The study was carried out in a Center for Dementia in collaboration with a university center. The primary outcomes were: (1) deprescribing inappropriate drugs through the Beers and STOPP&START criteria; (2) assessing duplicate drugs and the risk of iatrogenic damage due to drug–drug and drug–disease interactions. Overall, 69 men and 136 women (mean age 82.7 ± 7.4 years) were assessed. Of these, 91 patients were home care patients and 114 were outpatient. The average number of the drugs used in the sample was 9.4 drugs per patient; after the first visit and the consequent deprescribing process, the average dropped to 8.7 drugs per patient (p = 0.04). Overall, 74 potentially inappropriate drugs were used (36.1%). Of these, long half-life benzodiazepines (8.8%), non-steroidal anti-inflammatory drugs (3.4%), tricyclic antidepressants (3.4%), first-generation antihistamines (1.4%), anticholinergics (11.7%), antiplatelet drugs (i.e., ticlopidine) (1.4%), prokinetics in chronic use (1.4%), digoxin (>0.125 mg/day) (1.4%), antiarrhythmics (i.e., amiodarone) (0.97%), and α-blockers (1.9%) were included. The so-called “duplicate” drugs were overall 26 (12.7%). In total, ten potentially dangerous prescriptions were found for possible interactions (4.8%). We underline the importance of checking all the drugs taken periodically and discontinuing drugs with the lowest benefit-to-harm ratio and the lowest probability of adverse reactions due to withdrawal. Computer tools and adequately trained teams (doctors, nurses, and pharmacists) could identify, treat, and prevent possible drug interactions.