ABSTRACT: A wide gap exists between the evidence for acute demyelinating optic neuritis (ON) treatment and physicians’ clinical practice. The proposed research seeks to narrow this gap by designing an intervention to promote evidence-based, personalized treatment decisions. ON can cause permanent vision loss, but varia- tion exists in the visual prognosis. The Optic Neuritis Treatment Trial (ONTT) showed that, for the average pa- tient, intravenous corticosteroids did not improve visual outcomes, but may hasten recovery. This modest aver- age corticosteroids benefit comes with potential harms: medication side effects, and the inconvenience and cost of hospital admissions or outpatient infusions. Furthermore, the ONTT lacks modern neuro-axonal struc- tural outcomes and adequately powered prospective studies are lacking. Despite the variability in prognosis, modest benefits/risks of treatment, and limited understanding of corticosteroids effects on structural outcomes, providers treat nearly all patients with intensive corticosteroid therapy--adopting a “one-size-fits-all” approach to treatment. Personalizing ONTT data using multivariable risk stratified analyses and simulation analyses of neuroaxonal structural outcomes has great potential to inform individual treatment decisions. This strategy cir- cumvents the weaknesses of subgroup analyses, by examining the complex ways that multiple clinical factors shape an individual’s overall risks/benefits from treatment. However, information alone is not enough to change decision-making behavior. Perceptions about the treatment are important determinants of decision making be- havior. The objective of this proposal is to design a decision support intervention with personalized risk/benefit information that is informed by physician’s treatment perceptions to promote more nuanced, evidence-based ON treatment decisions. This objective will be addressed through 3 specific aims. (1) Establish the key treat- ment perceptions underlying physician’s ON decision making using a theory-based, mixed methods approach. (2) Develop models of the risks/benefits of corticosteroids at the individual level. (3) Use the results from Aims 1 and 2 to develop and pilot the usability, feasibility and acceptability of a decision support intervention. Dr. De Lott has a unique background in neuro-ophthalmology and health services research. She will develop addi- tional expertise in multivariable risk stratified analyses, simulation analyses, mixed methods, clinical trials, and design of decision support systems to become an independently-funded investigator committed to developing strategies to optimize evidence-based care. This proposal capitalizes on the unique strengths of the University of Michigan: Neurology and Ophthalmology Health Services Research Programs and the Institute for Health Care Policy and Innovation-among the largest academic health services research centers in the country. Dr. De Lott will be supported by an o...