Background:
Childhood exposure to adversity has been linked with later onset of severe mental illness. Obtaining good measures of exposure is made difficult by often poor recall quality, and the high cost, both in terms of clinical resources and time, required to collect information in quantities sufficient for valid scientific interrogation. Harvesting prospectively collected information stored on relevant state administrative registers (for example, mortality, morbidity, child protection and corrective services) offers an alternative means to assembling adversity constructs. Knowledge of the correspondence between register derived constructs and clinically derived ones is important to help assess the robustness of the former.
Aim:
To compare the values of 30 childhood adversity constructs derived from i) register records and ii) clinical note summaries.
Methods: For 189 children who were identified as having developed a severe mental illness, constructs of childhood adversity were derived using their: i) register records and ii) clinical case note summaries. Descriptive statistics summarized rates of identification and levels of agreement.
Results:
Rates of recording of adversities varied with generally higher rates identified in registers than clinical summaries. Constructs with the highest rates of clinical identification had the highest rates of agreement. Very few instances of contradiction between register and clinically derived constructs were observed.
Conclusion:
Constructs of childhood adversity derived from register records were observed to provide good agreement and little contradiction with corresponding constructs derived from clinical information. This supports confidence in the ability of register records to provide useful epidemiological measures of childhood adversity.