(PS10-84) Title: Creating an IRT Cross-walk Between the Promis-depression and -anxiety Scales and the SAPA Personality Inventory: Enabling Longitudinal Research on Mental Health in a Large American Sample

The Patient Reported Outcome Measurement Information System (PROMIS) is an ambitious project of the National Institutes of Health with the goal of improving the measurement of a wide range of physical and mental health outcomes. The initiative has created a publicly-available, psychometrically-robust computer adaptive testing system based on item response theory (IRT). The Synthetic Aperture for Personality Assessment (SAPA) is a data collection methodology that uses a Massively Missing Completely at Random design that delivers participants a small number of items randomly selected from a larger item pool (i >700), maximizing the breadth of constructs measured and utilizing measurements of items administered concurrently to build stable correlation matrices of many items. The SAPA project has collected data from millions of individuals worldwide over the past decade with the SAPA Personality Inventory (SPI) administered continuously across this time. In the current study, we leverage the large number of pairwise administrations between items on the SPI and PROMIS-Depression and -Anxiety (minimum pairwise administrations = 14,178) to create an IRT crosswalk that links the measures. Using a fixed-parameter calibration, we provide a psychometrically sound tool to convert the SPI items into PROMIS metrics, thereby enabling the analysis of depression and anxiety in a large sample of Americans across the past decade. For each crosswalk, we estimate unidimensionality and present item and test information curves to provide evidence for the validity of our findings. We share these cross-walked parameters and a script to enable researchers to convert the SAPA publicly available on the Harvard Dataverse into PROMIS metrics. We provide descriptive visualizations of PROMIS-Depression and -Anxiety across the past decade in the full sample of Americans to demonstrate the utility of the cross-walked SAPA dataset for studying the state of mental health in the United States.