This dissertation is comprised of three chapters, each of which contributes to the fields of health economics and psychology & economics. Two of the chapters investigate how choice architecture affects patients health care decisions. The third investigates the role of beliefs in the demand for health information. The underlying motivation for these studies stems from the observations that the proportion of HIV-infected persons in the United States who are undiagnosed has remained constant for over a decade, and those who are diagnosed are often diagnosed late in the course of their disease. Deeper understanding of how HIV test acceptance by patients depends on how the test is offered has the potential to decrease the frequency of missed opportunities for identifying infected persons.

Choice architecture has the potential to influence patients' health and health care deci- sions. In particular, decisions with significant but delayed consequences can be very sensitive to small, immediate costs and benefits. I investigate how small monetary incentives and de- fault test policies affect patient decision making with regard to HIV testing. I conduct and analyze the results of a field experiment that takes place in an urban emergency depart- ment (ED). In parallel with routine care, patients are approached by research assistants and offered HIV tests and questionnaires according to treatment assignments. In a factorial design, patients are randomly assigned to be offered HIV tests according to default scripts; they are also offered small monetary incentives and a questionnaire eliciting HIV-related risk behaviors. Patients are offered the questionnaire either before or after the test offer. Among those assigned to an early questionnaire, half are assigned to an additional question asking whether they would hypothetically accept an HIV test, a `foot-in-the-door' question (FITD).

In Chapter 1, "HIV Screening: To Test or Not to Test? It Depends on the Question," I examine three test defaults: traditional opt-in (test only those patients who request test- ing), opt-out (routine testing unless patients decline), and active-choice testing (patients are required to state whether they want to be tested). I find a test acceptance rate of 51.2% in the opt-in treatment. Active-choice and opt-out test schemes increased the proportion of patients who accepted HIV testing by clinically significant levels. Patients assigned to an active-choice test offer are 9.5 percentage points more likely to accept an HIV test; those assigned to an opt-out offer are 18.2 percentage points more likely than opt-in patients.

I take up the issue of monetary incentives in "Conditional Cash Incentives for HIV Test- ing." Patients are offered monetary incentives ($0, $1, $5, $10), which vary by ED zone (four zones) by day. I find that cash incentives of $5 and $10 increase test acceptance rates by 11.7 and 12.8 percentage points, respectively, from a baseline of 57.9% with no incentive. The $1 treatment assignment has no significant effect on overall test rates. It does, however, have a differential effect on high- and low-risk patients: patients reporting HIV risk factors are 4.3 percentage points more likely to test when offered $1 than when offered no incentive, and patients denying any risk factors are 9.6 percentage points less likely to accept testing when offered $1 than when offered no incentive. I find no difference in test rates between patients assigned to the FITD treatment and those in the early questionnaire treatment who were not asked the hypothetical question. Across defaults and monetary incentives, I observe an effect of being offered a questionnaire: patients assigned to either of Early or FITD questionnaires are 10.8 percentage points less likely to accept testing than those who are offered the test prior to being offered the questionnaire.

In "Perceptions and Misperceptions of HIV Transmission, Testing, and Treatment" I examine the relationship between beliefs regarding HIV transmission, testing, and treatment on subjects' testing behavior. I find that subjects grossly overestimate the probability of transmission for both real (e.g., injection drug use) and false (e.g., sharing a beverage) risk behaviors. Subjects also overestimate the prevalence of HIV. Despite these overestimates and self-reported risk behaviors, most consider themselves to be at no or very low risk of HIV infection, and few have ever tested for HIV. While these findings support both classical and behavioral interpretations (including psychological expected utility), these findings suggest that people's beliefs regarding HIV risk are biased and suggest that educating the public might be counterproductive, leading to more risky behaviors and lower testing rates.

Introduction: The short-term return visit rate among patients discharged from emergency departments (ED) is a quality metric and target for interventions. The ability to accurately identify which patients are more likely to revisit the ED could allow EDs and health systems to develop more focused interventions, but efforts to reduce revisits have not yet found success. Whether patients with a high number of ED visits are at increased risk of a return visit remains underexplored.

Methods: This was a population-based, retrospective, cohort study using administrative data from a large physician partnership. We included patients discharged from EDs from 80 hospitals in seven states from July 2014 – June 2016. We performed multivariable logistic regression of short-term return visits on patient, visit, hospital, and community characteristics. The primary outcome was the proportion of patients who had a return visit within 14 days of an index ED visit.

Results: Among 6,699,717 index visits, the overall risk of 14-day revisit was 12.6%. Frequent visitors accounted for 18.7% of all visits and 40.2% of all 14-day revisits. Frequent visitor status was associated with the highest odds of a revisit (odds ratio [OR] 3.06; 95% confidence interval [CI], 3.041 – 3.073). Other predictors of revisits were cellulitis (OR 2.131; 95% CI, 2.106 – 2.156), alcohol-related disorders (OR 1.579; 95%CI, 1.548 – 1.610), congestive heart failure (OR 1.175; 95% CI, 1.126 – 1.226), and public insurance (Medicaid OR 1.514; 95% CI, 1.501 – 1.528; Medicare OR 1.601; 95% CI, 1.583 – 1.620).

Conclusion: Previous ED use – even a single previous visit – was a stronger predictor of a return visit than any other patient, hospital, or community characteristic. Clinicians should consider previous ED use when considering treatment decisions and risk of return visit, as should stakeholders targeting patients at risk of a return visit.

Introduction: Emergency medical services (EMS) systems can become impacted by sudden surges that can occur throughout the day, as well as by natural disasters and the current pandemic. Because of this, emergency department crowding and ambulance “bunching,” or surges in ambulance-transported patients at receiving hospitals, can have a detrimental effect on patient care and financial implications for an EMS system. The Centralized Ambulance Destination Determination (CAD-D) project was initially created as a pilot project to look at the impact of an active, online base hospital physician and paramedic supervisor to direct patient destination and distribution, as a way to improve ambulance distribution, decrease surges at hospitals, and decrease diversion status. 

Methods: The project was initiated March 17, 2020, with a six-week baseline period; it had three additional study phases where the CAD-D was recommended (Phase 1), mandatory (Phase 2), and modified (Phase 3), respectively. We used coefficients of variation (CV) statistical analysis to measure the relative variability between datasets (eg, CAD-D phases), with a lower variation showing better and more even distribution across the different hospitals. We used analysis of co-variability for the CV to determine whether level loading was improved systemwide across the three phases against the baseline period. The primary outcomes of this study were the following: to determine the impact of ambulance distribution across a geographical area by using the CV; to determine whether there was a decrease in surge rates at the busiest hospital in this area; and the effects on diversion.

Results: We calculated the CV of all ratios and used them as a measure of EMS patient distribution among hospitals. Mean CV was lower in Phase 2 as compared to baseline (1.56 vs 0.80 P < 0.05), and to baseline and Phase 3 (1.56 vs. 0.93, P <0.05). A lower CV indicates better distribution across more hospitals, instead of the EMS transports bunching at a few hospitals. Furthermore, the proportion of surge events was shown to be lower between baseline and Phase 1 (1.43 vs 0.77, P <0.05), baseline and Phase 2 (1.43 vs. 0.33, P < 0.05), and baseline and Phase 3 (1.43 vs 0.42, P < 0.05). Diversion was shown to increase over the system as a whole, despite decreased diversion rates at the busiest hospital in the system. 

Conclusion: In this retrospective study, we found that ambulance distribution increased across the system with the implementation of CAD-D, leading to better level loading. The surge rates decreased at some of the most impacted hospitals, while the rates of hospitals going on diversion paradoxically increased overall. Specifically, the results of this study showed that there was an improvement when comparing the CAD-D implementation vs the baseline period for both the ambulance distribution across the system (level loading/CV), and for surge events at three of the busiest hospitals in the system.

Introduction: A crucial, yet subjective and non-evidence-based, decision for researchers is where to submit their original research manuscripts. The approach of submitting to journals in descending order of impact factor (IF) is a common but imperfect strategy. The validity of the IF as a measure of journal quality and significance is suspect, and a number of other journal impact scores have emerged, such that no one scale is universally accepted. Furthermore, practical considerations, such as likelihood of manuscript acceptance rates and times for decisions, may influence how authors prioritize journals. In this report, we sought to 1) review emergency medicine (EM) journal impact metrics, and 2) provide a comprehensive list of pertinent journal characteristics that may influence researchers’ choice of submission.

Methods: We systematically reviewed five impact metrics (IF, H Index, CiteScore, Source-Normalized Impact per Paper, and SCImago Journal Rank) and other relevant characteristics of 20 EM journals.

Results: We found good to excellent agreement in ordinal rankings of four of the journal impact metrics, as measured by the Spearman rank correlation coefficient. The median acceptance rate for original research manuscripts in the EM category was 25% (interquartile range [IQR] 18, 31%), and the median initial decision time was 33 days (IQR 18, 56 days). Fourteen EM journals (70%) accepted brief reports, and 15 (75%) accepted case reports/images.

Conclusion: We recommend replication, expansion, and formalization of this repository of information for EM investigators in a continuously updated, open-access forum sponsored by an independent organization.