Objective To compare the rate of detection of psychiatric illness in routine clinical practice, with the prevalence of mental illness established, using a dedicated screening programme at a regional neuroscience centre; and to assess if the screening programme had any enduring impact on routine clinical practice after its completion. Method Consecutive admissions to a neurology ward in the 3-month period before (n=160) and after (n=158) a dedicated neuropsychiatric screening program was carried out were identified. Case notes were then reviewed to establish if symptoms of mental illness were identified by the treating neurologists and if patients were referred for neuropsychiatric assessment. Rates of detection of neuropsychiatric problems and rates of referral for treatment were compared with those identified during the screening programme. Results In routine clinical practice, over two 3-month study periods, psychiatric symptoms were identified in 23.7% of patients and only 10.6% received neuropsychiatric interventions. This is much lower as compared to rates identified (51.3%) and treated (51.3%) during dedicated screening. Detection of mood symptoms decreased from 14.7% pre-screening to 3.8% in the post-screening period. Conclusion Rate of detection and treatment of neuropsychiatric problems remains low in neurology inpatients in routine clinical practice. Neuropsychiatric Screening is effective but does not have sustained effect once it stops. Hence we suggest active ongoing screening should be incorporated into routine practice.