Many NBER-affiliated researchers publish some of their findings in medical and other journals that preclude pre-publication distribution. This makes it impossible to include these papers in the NBER working paper series. This is a partial listing of recent papers in this category by NBER affiliates.
Regression to the Mean in the Medicare Hospital Readmissions Reduction Program
JAMA Internal Medicine 179(9), June 2019, pp. 1167–1173.
The Medicare Hospital Readmissions Reduction Program (HRRP), implemented in 2013, financially penalized hospitals with readmission rates above the national mean (deemed excess rates) for heart failure, heart attack, and pneumonia. Excess 30-day readmissions for these conditions have declined substantially in the hospitals that were initially penalized for high readmission rates. One potential explanation for this fact is that the policy incentivized these hospitals to improve care processes. However, an alternative explanation is the statistical phenomenon known as “regression to the mean,” which would predict that hospitals with rates farther from the mean in one period would be likely to fall closer to the mean in subsequent periods due to random chance. The researchers demonstrate that similar patterns of improvement in readmissions rates at hospitals with initially high rates occurred in settings with no financial incentives from the HRRP, such as for a diagnosis that was not evaluated under the program (chronic obstructive pulmonary disease), and during time periods before the program. The researchers conclude that most of the decline in readmission rates at hospitals that were initially penalized for high rates was attributable to regression to the mean, rather than to HRRP policy incentives.
Emergency Department Closures and Openings: Spillover Effects on Patient Outcomes in Bystander Hospitals
Health Affairs 38(9), September 2019, pp. 1496–1504.
When a hospital emergency department closes, the closest nearby hospitals with emergency departments (ED) can become overcrowded, particularly if they are already operating near capacity. An ED closing can also require patients to travel farther, potentially delaying treatment and increasing the severity of their condition. This study examines outcomes for heart attack patients at hospitals that were affected by an ED closure or opening. The sample includes Medicare beneficiaries who were treated for heart attacks at all EDs operating in the United States between 2001 and 2013: over 1 million patients across 3,720 hospitals. When an ED at a hospital with a high annual occupancy rate was affected by the closure of another ED that resulted in increased driving time of thirty minutes or more to the next-closest ED, the thirty-day readmission rates increased by 2.00 percentage points, one-year mortality increased by 2.39 percentage points, and the likelihood of receiving percutaneous coronary intervention (PCI) declined by 2.06 percentage points. On the other hand, hospitals that were affected by ED openings that resulted in decreased driving times of thirty minutes or more experienced reductions in thirty-day mortality and an increased likelihood of receiving PCI. However, ED openings and closings were generally not associated with changes in health outcomes at lower-occupancy hospitals or when the change in driving time was less than 30 minutes. The findings suggest that limited resources at high-occupancy hospitals make them particularly sensitive to changes in the availability of emergency care in neighboring communities, and that utilization as well as distance from neighboring ED’s can be important considerations when deciding whether to open or close an ED.
Do Dehydroepiandrosterone, Progesterone, and Testosterone Influence Women's Depression and Anxiety Levels? Evidence from Hair-Based Hormonal Measures of 2105 Rural Indian Women
Psychoneuroendocrinology, 109, November 2019, 104382.
Differences in sex hormone levels by gender are suspected to contribute to disparities in depression and anxiety, as suggested by both the higher prevalence of these conditions among women, and the abundance of sex-steroid receptors in brain areas relevant to these mental health problems. While this hypothesis has been supported by animal research, research in human populations has been less conclusive. This study examines the correlation between hair concentrations of sex hormones and symptoms of depression and anxiety in a sample of 2,105 rural Indian women aged 18–85 years, from a financial survey of Indian households spanning 876 villages. The researchers find that increased psychological distress, as measured by a screening questionnaire, was associated with higher levels of testosterone and lower levels of dehydrepiandrosterone (DHEA). Progesterone levels did not show a clear association with either depressive or anxiety symptoms. DHEA results were not affected by omitting data from those with non-detectible values, rather than using imputed values, but testosterone results were. When omitting non-detectible values, testosterone more weakly predicted psychological distress, and trended negative in its relation to depressive symptoms. The study suggests a potential protective effect of higher DHEA levels, and the testosterone findings have potentially important implications for the interpretation of studies in which non-detectable values are excluded.