OBJECTIVES: In pharmaceutical outcomes research using healthcare claims databases, periods during which patients are hospitalized have constituted “black holes”, as these databases do not contain any information on pharmacotherapy received in hospital.  While admission-level databases provide such information, they lack information on pharmacotherapy received outside of hospital.  Recently, it has become possible to link these two types of databases.  In this study, we explored their potential value to outcomes research, focusing attention on second-generation antipsychotic (SGA) treatment before, during, and after hospitalization for schizophrenia. METHODS: Using a linked inpatient/outpatient database, we identified all adults with ≥1 admissions for schizophrenia between January 1, 2001 and September 30, 2008.  Focusing on each patient’s first admission, we compiled all healthcare claims during the 6-month periods preceding and following hospitalization.  As our interest was in the use of SGAs, our scope was limited to patients with evidence of receipt of oral ziprasidone, aripiprazole, or quetiapine (“study agents”) immediately preceding hospital discharge.  We then examined receipt of these agents during the 6-month periods preceding hospitalization and following hospital discharge based on outpatient pharmacy claims; receipt of study agents in hospital was examined using admission-level data.  Adherence with study agents following hospital discharge was assessed using proportion of days covered (PDC); patients were deemed nonadherent if PDC fell below 80%. RESULTS: A total of 43 patients were identified who met all study entry criteria.  Twenty-four patients (56%) had evidence of receipt of a study agent in the period preceding hospitalization.  While all patients had evidence of receipt of study agents following hospital discharge, only 12% were adherent at 6 months.   CONCLUSIONS: Linked inpatient/outpatient databases are a promising avenue for future pharmaceutical outcomes research, as they may greatly expand understanding of the complete chronology of pharmacotherapy‑‑and associated outcomes--for many disease conditions.