How complete is the information on preadmission psychotropic medications in inpatients with dementia? A comparison of hospital medical records with dispensing data

Abstract

Reliable information on preadmission medications is essential for inpatients with dementia, but its quality has hardly been evaluated. We assessed the completeness of information and factors associated with incomplete recording. We compared preadmission medications recorded in hospital electronic medical records (EMRs) with community-pharmacy dispensations in hospitalizations with discharge code for dementia at the University Hospital of Udine, Italy, 2012-2014. We calculated: (a) prevalence of omissions (dispensed medication not recorded in EMRs), additions (medication recorded in EMRs not dispensed), and discrepancies (any omission or addition); (b) multivariable logistic regression odds ratio, with 95% confidence interval (95% CI), of ≥1 omission. Among 2,777 hospitalizations, 86.1% had ≥1 discrepancy for any medication (Kappa 0.10) and 33.4% for psychotropics. When psychotropics were recorded in EMR, antipsychotics were added in 71.9% (antidepressants: 29.2%, antidementia agents: 48.2%); when dispensed, antipsychotics were omitted in 54.4% (antidepressants: 52.7%, antidementia agents: 41.5%). Omissions were 92% and twice more likely in patients taking 5 to 9 and ≥10 medications (vs. 0 to 4), 17% in patients with psychiatric disturbances (vs. none), and 41% with emergency admission (vs. planned). Psychotropics, commonly used in dementia, were often incompletely recorded. To enhance information completeness, both EMRs and dispensations should be used.