Difference-in-differences Estimate Research Articles

High-speed mobile networks and police repression during the COVID-19 pandemic: the case of Nigeria

PurposeGovernment repression against civilians while enforcing COVID-19 related lockdowns was widely reported in Africa. At the same time, many have claimed that high-speed (4G) mobile network proliferation provide an accountability mechanism that may constrain police abuses. This study focused on Nigeria to examine (1) the effect of COVID-19 lockdowns on police repression and (2) whether widespread high-speed mobile data networks constrain police repression.Design/methodology/approachUsing data from the Armed Conflict Location and Event Database (ACLED) and the Mobile Coverage Database, this study used difference-in-differences (DID) and triple difference (DDD) estimation on a sample of 423,925 observations (local government area-days) between January 1, 2019 and June 30, 2020 to estimate the causal effects of COVID-19 lockdowns and high-speed (4G) mobile data on police repression.FindingsLockdowns increased certain forms of police repression in areas with substantial high-speed (4G) mobile networks. Separate from the lockdowns, widespread 4G network increased police repression even without lockdowns.Research limitations/implicationsProliferation of high-speed mobile networks in Nigeria appears to facilitate, rather than constrain, police repression. It is possible that high-speed mobile data networks allow police to detect and repress citizen behaviors, rather than permitting citizens to correct repressive police behaviors.Originality/valueAlthough many studies have explored the COVID-19 pandemic and police behavior in Western countries, only a few have examined its effects in states with even more troubled policing institutions, including those in sub-Saharan Africa, using DID and DDD estimation.

Association between state Medicaid policies and accrual of Black and Hispanic patients to cancer clinical trials.

1510 Background: The U.S. Medicaid health insurance program may play a key role in mitigating racial and ethnic disparities in cancer clinical trial enrollment, given Medicaid’s disproportionate representation of these populations and its coverage of the routine costs of trial participation in some states. We sought to evaluate the effect of Medicaid expansion under the Affordable Care Act (ACA) on changes in the accrual of racial and ethnic minority patients in cancer clinical trials, across all states and specifically those mandating Medicaid coverage of routine trial costs. Methods: The study used anonymous enrollment data from Medidata Rave, a large commercial clinical trial data collection platform. We conducted a retrospective difference-in-differences (DID) analysis to examine accrual demographics in states that expanded Medicaid eligibility under the ACA compared with non-expansion states, before and after expansion; and assessed effect modification based on state-mandated Medicaid coverage of routine trial costs. Participants included adults enrolled in interventional cancer clinical trials from 2012-2019, located in U.S. hospital referral regions with >5% Black or Hispanic population. The primary outcome was participant race/ethnicity, defined dichotomously as Black or Hispanic (vs non-Black, non-Hispanic). We used ordinary least squares regression to model participant race/ethnicity as a function of (1) an indicator for trial enrollment in an expansion state after policy implementation, (2) state fixed effects, and (3) year fixed effects, with standard errors clustered at the state level. To assess effect modification, we stratified the main model across states with and without coverage mandates of routine trial costs. Results: The study included 75,700 participants across 1,409 clinical trials and 351 sites. In expansion states, the proportion of participants who identified as Black and/or Hispanic increased from 13.4% before expansion to 13.8% after expansion (0.4 percentage-point [PP] change, 95% CI -0.8 to 1.7). In non-expansion states, this proportion decreased from 17.0% to 15.2% (-1.7 PP change, 95% CI -3.7 to 0.2). These trends yielded a non-significant DID estimate of 2.0 PP (95% CI -1.0 to 5.0). In stratified analyses, Medicaid expansion was associated with a 4.5 PP (95% CI 0.4 to 8.6) increase in the enrollment of Black and/or Hispanic participants in states with coverage mandates, but not in states without coverage mandates (1.1 PP, 95% CI -2.6 to 4.8). Conclusions: Medicaid expansion was associated with a significant increase in the proportion of Black and Hispanic oncology trial participants in states that mandated Medicaid coverage of the routine costs of trial participation, but not overall across all states. These findings suggest that Medicaid policies have the potential to advance equitable enrollment in cancer clinical trials.

Medicaid enrollment changes among U.S. adult survivors of childhood cancer following medicaid expansion: A report from the Childhood Cancer Survivor Study (CCSS).

10011 Background: Little is known about whether Medicaid expansion under the Affordable Care Act (ACA) affected insurance coverage among adult survivors of childhood cancer, a population at high-risk for poor health outcome. We addressed this gap by evaluating the association between ACA Medicaid expansion and Medicaid enrollment among participants in the Childhood Cancer Survivor Study (CCSS). Methods: The CCSS cohort of 5-year survivors of childhood cancer was linked to administrative Medicaid insurance data in 2010-2016. We identified 13,895 adult survivors (aged 18-64 years) diagnosed with cancer under age 21 and between 1970 and 1999. Outcomes included (1) percentage with any Medicaid enrollment by year; and (2) Medicaid-covered days (number of days when a survivor was enrolled in Medicaid) during each year. Multivariable difference-in-differences (DD) models were used to examine outcome changes pre vs. post ACA Medicaid expansion, in expansion- vs. non-expansion states, adjusting for age, sex, race/ethnicity, income, education, and chronic conditions. Multivariable models were conducted overall and then stratified by cancer type, race/ethnicity, income, and education. Results: Medicaid enrollment rates increased more in expansion states (17.6% pre-expansion to 24.1% post-expansion) than non-expansion states (16.4% to 16.9%), leading to a net increase in enrollment of 6.6 percentage points (ppt; 95% CI = 5.5-7.7) in the multivariable DD model. Multivariable DD model showed a net increase of 18.4 days (95% CI = 13.8-23.1) in Medicaid-covered days in expansion states relative to non-expansion states. The expansion-associated increase in Medicaid enrollment rates was greatest among survivors of leukemia (multivariable DD estimate: 8.9 ppt, 95% CI = 6.9-11.0) and non-Hodgkin lymphoma (8.0 ppt, 95% CI = 5.0-10.9). Greater increases in Medicaid enrollment were seen in non-Hispanic Black (13.5 ppt, 95% CI = 8.0-19.1) and Hispanic survivors (15.8 ppt, 95% CI = 10.6-21.0) than non-Hispanic White peers (5.1 ppt; 95% CI = 4.0-6.2); in survivors with < $20K household income (11.7 ppt, 95% CI = 8.9-14.4) compared to those with ≥$60K household income (2.9 ppt, 95% CI = 1.6-4.2); and in survivors with high school or lower education (9.3 ppt, 95% CI = 6.9-11.8) compared to those with a college or higher degree (3.9 ppt, 95% CI = 2.6-5.3). Similar patterns were observed across survivor subgroups when examining Medicaid-covered days. Conclusions: We provide the first evidence on increased Medicaid enrollment and longer coverage duration among adult survivors of childhood cancer following Medicaid expansion, with greater increases seen among survivors of underrepresented racial/ethnic populations, those with low socioeconomic status, and high medical need survivors.

The Long-Term Impact of Famine Experience on Harvest Losses

Approximately one-third of the global food supply is lost or wasted each year. Given that the harvesting process is the initial stage following food production, minimizing losses in this crucial phase holds paramount significance in augmenting the food supply and ensuring food security. The 1959–1961 famine in China was one of the most catastrophic events in history and had long-term effects on human beings, particularly farmers. This paper aims to provide a new perspective on the variations in harvest losses across age cohorts by examining the impact of famine experiences. Using survey data from the 2016 Postproduction Food Loss and Waste Survey conducted by China Agricultural University and the Rural Economic Research Center, which involved 3538 farming households across 28 provinces, we construct a cohort difference-in-difference (DID) model to investigate the impact of famine experience on household harvest losses. The standard cohort DID estimation results indicate that in areas with severe famine, a 1% increase in excess mortality would reduce the rate of harvest loss by 3%, suggesting that farmers who have experienced extreme famine have a deeper memory of the event, which subsequently helps them reduce harvest losses. Moreover, the results of the heterogeneity test reveal that the more serious the famine that the household head experienced in early life was, the less harvest losses there were, particularly for those who were adolescents during the famine. The findings elucidate the importance of historical events in shaping current behaviors and contribute to a better understanding of the variation in harvest losses across age cohorts.

Contribution of home garden vegetables on reducing stunting among 6- to 23-month-old children in South Tigray, northern Ethiopia.

The study was conducted to analyze the contribution of home garden vegetables in reducing stunting among 6- to 23-month-old children from South Tigray, Northern Ethiopia. The quasi-experimental study design was used. Multistage sampling technique was used to select the districts and study communities. A total of 94 purposively selected vegetable producer (intervention) households and 260 randomly selected non-producer (control) households were included in the study (1:3 ratio). The recumbent length of children was measured using horizontal wooden board to the nearest 0.1 cm. The length-for-age Z-scores were computed using WHO-Anthro 2006 software. Propensity score-matching and difference-in-difference (DID) estimates were used to analyze data using STATA software version 12. Prevalence of child stunting was 19.8 (12.7-29.4) and 21.1 (16.4-26.7)% (baseline) and 43.5 (33.5-54.1) and 46.5 (45.7-47.2)% (end line) among intervention and control groups, respectively. Child stunting was higher for boys and older children from both intervention and control households. DID estimation revealed that there was no significant difference in child stunting between intervention and control households (DID = 1.7, p = .604). However, there was an intervention effect of -0.5, 2.5, and 1.7% in the prevalence of child stunting among females, males, and both sexes, respectively. Vegetable production as an intervention strategy reduced the prevalence of stunting in children aged 6-23 months. However, vegetable production needs to be well integrated with other nutrition-sensitive interventions to realize the objective of reducing child stunting.

The validity of causal claims with repeated measures designs: A within-study comparison evaluation of differences-in-differences and the comparative interrupted time series.

Modern policies are commonly evaluated not with randomized experiments but with repeated measures designs like difference-in-differences (DID) and the comparative interrupted time series (CITS). The key benefit of these designs is that they control for unobserved confounders that are fixed over time. However, DID and CITS designs only result in unbiased impact estimates when the model assumptions are consistent with the data at hand. In this paper, we empirically test whether the assumptions of repeated measures designs are met in field settings. Using a within-study comparison design, we compare experimental estimates of the impact of patient-directed care on medical expenditures to non-experimental DID and CITS estimates for the same target population and outcome. Our data come from a multi-site experiment that includes participants receiving Medicaid in Arkansas, Florida, and New Jersey. We present summary measures of repeated measures bias across three states, four comparison groups, two model specifications, and two outcomes. We find that, on average, bias resulting from repeated measures designs are very close to zero (less than 0.01 standard deviations; SDs). Further, we find that comparison groups which have pre-treatment trends that are visibly parallel to the treatment group result in less bias than those with visibly divergent trends. However, CITS models that control for baseline trends produced slightly more bias and were less precise than DID models that only control for baseline means. Overall, we offer optimistic evidence in favor of repeated measures designs when randomization is not feasible.

Can e-commerce reduce urban CO2 emissions? Evidence from National E-commerce Demonstration Cities policy in China.

The existing studies ignore the potential impact of e-commerce on CO2 emissions. Based on panel data of 284 cities at the prefecture level and above in China from 2003 to 2019, we regard National E-commerce Demonstration Cities (NEDC) policy as a quasi-natural experiment and empirically examine the impact and mechanism of NEDC policy on urban CO2 emissions by using multi-period difference-in-difference (DID) model and mediating effect model. The results show that, under the premise of satisfying parallel trends, the NEDC policy significantly reduces urban CO2 emissions. Compared with the non-pilot cities, the NEDC policy leads to decreases in urban CO2 emissions by 0.83%. The result still holds after a series of robustness tests, which include placebo test, propensity score matching (PSM) DID estimation, and instrumental variable estimation. The NEDC policy reduces urban CO2 emissions through promoting industrial structure upgrading, green technology innovation, and economic agglomeration. Heterogeneity analysis shows that the reduction effect of NEDC policy on urban CO2 emissions is stronger in non-western cities, larger scale cities, and resource-based cities.

Controlling time-varying confounding in difference-in-differences studies using the time-varying treatments framework

This article clarifies how the biostatistical literature on time-varying treatments (TVT) can provide tools for dealing with time-varying confounding in difference-in-differences (DiD) studies. I use a simulation study to compare the bias and standard error of inverse probability weighting estimators from the TVT framework, a DiD framework, and hybrid approaches that combine ideas from both frameworks. I simulated longitudinal data with treatment effect heterogeneity over multiple time points using linear and logistic models. Simulation settings looked at both time-invariant confounders and time-varying confounders affected by prior treatment. Estimators that combined ideas from both frameworks had lower bias than standard TVT and DiD estimators when assumptions were unmet. The TVT framework provides estimation tools that can complement DiD tools in a wide range of applied settings. It also provides alternate estimands for consideration in policy settings.

The impact of delayed school start time on adolescent beverage consumption, findings from the START study

ObjectivesWe sought to examine the effects of high school start time delay, a proven sleep-promoting intervention, on sugary beverage (SB) consumption among U.S. adolescents. MethodsIn the Spring of 2016 (baseline), the START study recruited 2134 ninth grade students who were enrolled high schools in the Twin Cities, MN metropolitan area. These participants were surveyed again in their 10th and 11th grade years, in Spring 2017 and 2018 (follow-ups 1 and 2). All five high schools started early (7:30 or 7:45 a.m.) at baseline. By follow-up 1, two “policy change” schools shifted their start times later (8:20 or 8:50 a.m.) and maintained these later start times through follow-up 2 while three “comparison schools” retained an early start time at all time points. Generalized estimating equations using a negative binomial distribution were used to obtain estimates of the number of sugary beverages consumed per day at each wave as well as the difference in difference (DiD) estimates between baseline and each follow-up period comparing policy change to comparison schools. ResultsMean baseline sugary beverage consumption was 0.9 (1.5) beverages per day in policy change schools and 1.2 (1.7) beverages per day in the comparison schools. While there was no evidence of impact of start time change on total SB consumption, DiD estimates revealed a small decrease in the number of caffeinated sugary beverages consumed between baseline and the second follow-up period in students attending the policy change schools relative to comparison schools in both crude (0.11/day reduction, p-value = 0.048) and adjusted analyses (0.11/day reduction, p-value = 0.028). ConclusionAlthough the differences in this study were quite modest, a population-wide reduction in sugary beverage consumption could have public health benefit.

Setting goal difficulty in monetary incentives to physicians: evidence from an online health knowledge-sharing platform

PurposePrior studies on the impact of incentives on physicians’ online participation mainly focused on different incentives while ignoring the difficulty of setting monetary incentives efficiently. Based on goal-setting theory, the current research examines the relationship between incentives with goals of varying difficulty and professional health knowledge sharing (PHKS) in online health knowledge-sharing platforms (OHKSPs).Design/methodology/approachFour field experiments with different monetary incentives were conducted by one of China’s largest OHKSPs, with whom the researchers cooperated in data collection. Monthly panel data on 10,584 physicians were collected from September 2018 to December 2019. There were 9,376 physicians in the treatment group and 1,208 in the control group. The authors used a difference-in-difference (DID) model to explore the research question based on the same control group and the Chow test with seemingly unrelated estimation (sureg) to compare regression coefficients between four groups. Several robustness checks were performed to validate the main results, including a relative time model, multiple falsification tests and a DID estimation using the propensity score matching method.FindingsThe results show that the monetary incentive significantly positively affected the volume of physicians’ PHKS directly with negative spillover to the duration of physicians’ PHKS. Moreover, the positive effect of incentives with higher difficulty on the volume of physicians’ PHKS was significantly smaller than that of incentives with low difficulty. Finally, professional title had a positive moderating effect on the volume of goal difficulty setting and did not significantly moderate the effect on the duration of physicians’ PHKS.Research limitations/implicationsSome limitations of this study are: firstly, because the field experiments were enterprise benefit oriented, the treatment and control groups were not balanced. Secondly, the experiments for different incentive measures were relatively similar, making it challenging to validate a causal effect. Finally, more consideration should be given to the strategy for setting hierarchical incentives in future research.Originality/valueThe research indicates that monetary incentives have a bilateral effect on PHKS, i.e. a positive direct effect on the volume of physicians’ contributions and a negative spillover effect on the duration of physicians’ PHKS. The professional titles of physicians also moderate such bilateral switches of PHKS. Furthermore, when a physician’s energy is limited, the goal difficulty setting of the incentive mechanism tends to be low. The more difficult the incentives are, the more inefficient the effects on physicians’ PHKS will be.

Practical operation and theoretical basis of difference-in-difference regression in science of science: The comparative trial on the scientific performance of Nobel laureates versus their coauthors

Abstract Purpose In recent decades, with the availability of large-scale scientific corpus datasets, difference-in-difference (DID) is increasingly used in the science of science and bibliometrics studies. DID method outputs the unbiased estimation on condition that several hypotheses hold, especially the common trend assumption. In this paper, we gave a systematic demonstration of DID in the science of science, and the potential ways to improve the accuracy of DID method. Design/methodology/approach At first, we reviewed the statistical assumptions, the model specification, and the application procedures of DID method. Second, to improve the necessary assumptions before conducting DID regression and the accuracy of estimation, we introduced some matching techniques serving as the pre-selecting step for DID design by matching control individuals who are equivalent to those treated ones on observational variables before the intervention. Lastly, we performed a case study to estimate the effects of prizewinning on the scientific performance of Nobel laureates, by comparing the yearly citation impact after the prizewinning year between Nobel laureates and their prizewinning-work coauthors. Findings We introduced the procedures to conduct a DID estimation and demonstrated the effectiveness to use matching method to improve the results. As a case study, we found that there are no significant increases in citations for Nobel laureates compared to their prizewinning coauthors. Research limitations This study ignored the rigorous mathematical deduction parts of DID, while focused on the practical parts. Practical implications This work gives experimental practice and potential guidelines to use DID method in science of science and bibliometrics studies. Originality/value This study gains insights into the usage of econometric tools in science of science.

P851 Impact of an integrated model of care for inflammatory bowel disease on direct healthcare costs: A population-based matched cohort study from Saskatchewan, Canada

Abstract Background Integrated models of care (IMC) for IBD lead to reduced hospitalizations, surgeries, comorbidities, and overall improved outcomes. There are limited studies assessing the impact of IMC on direct healthcare costs. We aimed to estimate the impact of exposure to an IMC on direct healthcare costs among individuals diagnosed with IBD. Methods We conducted a quasi-experimental difference-in-difference (DID) cost analysis using administrative health data from Saskatchewan, Canada. We included individuals ≥18 years old meeting a validated administrative IBD case definition between January 2009 and March 2015. IBD cases were classified as exposed or non-exposed to the Saskatchewan IMC and required to have three years of healthcare coverage after baseline (i.e., first visit with an IMC [exposed] or a non-IMC [non-exposed] gastroenterologist). The direct healthcare costs were derived from the administrative databases, adjusted to 2014/15 Canadian dollars, and categorized into hospitalizations, physician visits, and medication claim costs (i.e., immunomodulators [IMM], biologics [BIOL], aminosalicylates[5-ASA]). Propensity scores (PS) were calculated based on healthcare utilization and comorbidities in the 12 months before baseline. Cases (exposed) were matched 1:5 controls (non-exposed) based on PS and disease duration. DID estimators were determined for each cost category using mixed linear regression models including age, sex, disease type, and area of residence as covariables. Results In total, 2905 IBD cases were included in the study, 597 exposed and 2308 non-exposed individuals (Figure); the majority were females (52%), lived in urban areas (74%), and had Crohn’s Disease (61%). The mean age was 44.6 years (SD=15.5) and the average disease duration at baseline was 5.7 years (SD=5). In comparison to the non-exposed group, the costs of physician visits (IBD-related [DID= $-139, 95%CI -209 to -68], IBD-specific [DID=$-155, 95%CI -212 to -97], and specialty visits [DID=$-135, 95%CI -201 to -69]); and IMM dispensations (DID=$-18, 95%CI -35 to -2) were lower in the exposed group. Conversely, we identified higher total healthcare costs (DID=$1707, 95%CI 369 to 3044); IBD-related medication (DID=$2341, 95%CI 1484 to 3197) and BIOL (DID= $2467, 95%CI 1605 to 3328) dispensations costs in the exposed group. There were no other statistically significant differences between the groups. Conclusion We identified lower physician visit and IMM costs in the exposed group than in the non-exposed one. Total healthcare costs were higher among the exposed group, largely driven by higher BIOL costs. These results highlight the potential cost impact of IMC for IBD and the need for cost-effectiveness studies of IMC to further assess value of this care model.

SUPPLEMENTAL PAYMENTS AND NURSING HOME STAFFING: A CASE STUDY OF INDIANA

Abstract Adequate reimbursements are key to improving quality in nursing homes (NHs), but Medicaid payments remain low. Indiana implemented Medicaid supplemental payment program to non-state government (NSGOs) owned NHs to boost Medicaid payment rates. Under the program, NSGOs contribute the state-share of the supplemental payments through intergovernmental transfers allowing the state to claim federal matching funds. The supplemental payments are distributed between NSGOs and NHs but little is known about whether NHs invested the supplemental payments to improve staffing levels. In this study, we examine whether NHs receiving supplemental payments, proxied by NSGO-ownership, have higher staffing hours per resident day (HPRD) for registered nurse (RN), licensed practical nurse (LPN), and certified nurse aide (CNA) compared with NHs without supplemental payments. We use facility-year data on NSGO ownership (State of Indiana), staffing levels (Nursing Home Compare), and facility and resident characteristics (LTCfocus.org) on 500 NHs from 2009 to 2017. We use difference-in-difference (DD) regressions with year and facility fixed effects to estimate the potentially causal effect of NSGO ownership on NH staffing controlling for facility and resident characteristics. DD estimates suggest that NSGO-ownership is associated with decreased staffing level across all three types of staffing. More specifically, RN staffing decreased by 0.019 HPRD (p< 0.05), LPN staffing decreased by 0.041 HPRD (p< 0,05), and CNA staffing decreased by 0.089 HPRD (p< 0.001). Policymakers should increase the transparency of supplemental payment programs and audit the expenditures to ensure that supplemental payments are tied to quality improvement.

Prevention of 90-day inpatient detoxification readmission for opioid use disorder by a community-based life-changing individualized medically assisted evidence-based treatment (C.L.I.M.B.) program: A quasi-experimental study.

Evidence for community-based strategies to reduce inpatient detoxification readmission for opioid use disorder (OUD) is scant. A pilot program was designed to provide individualized structured treatment plans, including addressing prolonged withdrawal symptoms, family/systems assessment, and contingency management, to reduce readmission after the index inpatient detoxification. A non-randomized quasi-experimental design was used to compare the pilot facilities (treatment) and comparison facilities before and after the program started, i.e., a simple difference-in-differences (DID) strategy. Adults 18 years and older who met the Diagnostic and Statistical Manual of Mental Disorders version 5 criteria for OUD and had an inpatient detoxification admission at any OUD treatment facility in two study periods between 7/2016 and 3/2020 were included. Readmission for inpatient detoxification in 90-days after the index stay was the primary outcome, and partial hospitalization, intensive outpatient care, outpatient services, and medications for OUD were the secondary outcomes. Six statistical estimation methods were used to triangulate evidence and adjust for potential confounding factors between treatment and comparison groups. A total of 2,320 unique patients in the pilot and comparison facilities with 2,443 index inpatient detoxification admissions in the pre- and post-periods were included. Compared with patients in comparison facilities, patients in the C.L.I.M.B. facilities had higher readmission in the pre-period (unadjusted readmission 17.0% vs. 10.6%), but similar rates in the post-period (12.3% vs. 10.6%) after the implementation of the pilot program. For 90-day readmission, all DID estimates were not statistically significant (adjusted estimates ranged from 6 to 9 percentage points difference favoring the C.L.I.M.B. program). There was no significant improvement in the secondary outcomes of utilizations in lower level of care and medications for OUD in C.L.I.M.B. facilities. We found a reduction in readmission in the pilot facilities between the two periods, but the results were not statistically significant compared with the comparison facilities and the utilization of lower level of care services remained low. Even though providers in the pilot OUD treatment facilities actively worked with health plans to standardize care for patients with OUD, more strategies are needed to improve treatment engagement and retention after an inpatient detoxification.

The Nonimpact of Opportunity Zones on Home and Business Lending

Opportunity Zones (OZs) promised to stimulate investment in over 8,500 “distressed” neighborhoods. Have OZs increased neighborhood investment and, if so, what types of neighborhoods have benefitted? This study uses a difference-in-differences (DID) design to compare small business and residential lending outcomes in designated OZs with areas that were eligible but not designated. Census tracts are stratified by pretreatment social and economic indicators of distress to search for heterogeneity in effects by neighborhood type. An event study framework is used to check the parallel trends assumption and census tracts are then matched to improve the counterfactual. Finally, an adjusted interrupted time series (AITS) analysis is introduced to further evaluate differences in outcome indicator levels and trends between target and control neighborhoods pre and post OZ. DID and AITS estimates suggest that OZs have had no statistically significant effects on business or residential loan growth. Heterogeneity modeling confirms a noneffect across neighborhood distress type. In conclusion, study limitations and future extensions for both policy and research are discussed.

The impact of increasing expenditure on National Essential Public Health Services on the medical costs of hypertension in China: A difference-in-difference analysis.

The prevention and control of hypertension should be an effective way to reduce deaths and it has been a high priority in China. In 2013, the Chinese government increased the subsidy standard for the National Essential Public Health Services Package (NEPHSP) from RMB 15 to RMB 30 per person, which was expected to cover 70 million hypertensions. This study explored the influence of increasing NEPHSP subsidy on outpatient and inpatient expenditure among patients with hypertension. Data were mined from the 2011-2015 Harmonized China Health and Retirement Longitudinal Study. The study sample included 3192 hypertensive patients who were not lost to follow-up from 2011 to 2015. Hypertensive patients who covered by NEPHSP from 2011 to 2015 were defined as the treatment group, otherwise defined as the comparison group. The policy intervention was the increase of NEPHSP subsidy in 2013, and the years before and after 2013 were respectively considered as pre- (2011) and post-intervention (2015). The primary outcomes variables were the outpatient and inpatient expenditure of patients with hypertension, based on direct spending of outpatients and inpatients separately reported by patients with hypertension. Using propensity score matching (PSM) to match the individual characteristics of hypertension in the treatment group and the comparison group, difference-in-differences (DID) were used to analyze the outcomes. The patients with hypertension' outpatient and inpatient expenditure patterns in the treatment and control group show an increasing trend from 2011 to 2015. After PSM, of the 1 956 hypertensive participants, 369 covered by the NEPHSP before and after 2013. A DID estimate of the increased NEPHSP subsidy was associated with a significant decrease of 1 251.35 RMB (t = 2.13, P = 0.034) in hypertension related inpatient expenditure, no significant change (t = 0.61, P = 0.544) among outpatient expenditure. The NEPHSP may reduce inpatient expenditure among hypertension. Further strengthening of the NEPHSP may reduce their burden.

The impact of Illinois’ comprehensive handheld phone ban on talking on handheld and handsfree cellphones while driving

Introduction: Distracted driving has been linked to multiple driving decrements and is responsible for thousands of motor-vehicle fatalities annually. Most U.S. states have enacted restrictions on cellphone use while driving, the strictest of which prohibit any manual operation of a cellphone while driving. Illinois enacted such a law in 2014. To better understand how this law affected cellphone behaviors while driving, associations between Illinois’ handheld phone ban and self-reported talking on handheld, handsfree, and any cellphone (handheld or handsfree) while driving were estimated. Methods: Data from annual administrations of the Traffic Safety Culture Index from 2012–2017 in Illinois and a set of control states were leveraged. The data were cast into a difference-in-differences (DID) modeling framework, which compared Illinois to control states in terms of pre- to post-intervention changes in the proportion of drivers who self-reported the three outcomes. Separate models for each outcome were fit, and additional models were fit to the subset of drivers who talk on cellphones while driving. Results: In Illinois, the pre- to post-intervention decrease in the drivers’ probability of self-reporting talking on a handheld phone was significantly more extreme than that of drivers in control states (DID estimate −0.22; 95% CI −0.31, −0.13). Among drivers who talk on cellphones while driving, those in Illinois exhibited a more extreme increase in the probability of talking on a handsfree phone while driving than those control states (DID estimate 0.13; 95% CI 0.03, 0.23). Conclusions: These results suggest that Illinois’ handheld phone ban reduced talking on handheld phones while driving among study participants. They also corroborate the hypothesis that the ban promoted substitution from handheld to handsfree phones among drivers who talk on the phone while driving. Practical Applications: These findings should encourage other states to enact comprehensive handheld phone bans to improve traffic safety.

Recreational marijuana laws and the misuse of prescription opioids: Evidence from National Survey on Drug Use and Health microdata.

Several studies have concluded that legalizing medical marijuana can reduce deaths from opioid overdoses. Drawing on micro data from the National Survey on Drug Use and Health, a survey uniquely suited to assessing patterns of substance use, we examine the relationship between recreational marijuana laws (RMLs) and the misuse of prescription opioids. Using a standard difference-in-differences (DD) regression model, we find that RML adoption reduces the likelihood of frequently misusing prescription opioids such as OxyContin, Percocet, and Vicodin. However, using a two-stage procedure designed to account for staggered treatment and dynamic effects, the DD estimate of relationship between RML adoption and the likelihood of frequently misusing prescription opioids becomes positive. Although event study estimates suggest that RML adoption leads to a decrease in the frequency of prescription opioid abuse, this effect appears to dissipate after only 2 or 3years.

Is it just a cure? Re-evaluating the effects of prenatal exposure to the Great Chinese Famine on the risk of infectious diseases in adulthood

BackgroundVery limited studies focused on the early-life adversities on infectious diseases. Taking the Great Chinese famine as a natural experiment, this study re-evaluated the long-term effect of prenatal famine exposure on infectious diseases by using nationally representative data. MethodsUsing difference-in-difference (DID) models, we analyzed 215,216 adults who participated in the Second National Sample Survey on Disability in 2006 across 734 counties of China to detect the effects of prenatal exposure to the Great Chinese Famine on the risk of infectious diseases in adulthood. Infectious diseases were ascertained by using the combination of self-reports or family members’ reports and on-site medical diagnosis by experienced specialists, and the severity of famine was measured by the cohort size shrinkage index (CSSI) at the county level. ResultsAll DID estimates of the effects of famine on the probability of infectious diseases were insignificant, with a coefficient of 0.0007 (−0.0024, 0.0026) for all participants and coefficients of 0.0001 (−0.0041, 0.0043) and −0.0002 (−0.0036, 0.0033) for males and females, respectively. That is, the famine cohort dwelling in regions with a greater intensity of famine had similar levels of infectious disease risks than the cohorts with post-famine prenatal exposure experience in less affected famine regions. Furthermore, there were no significant famine and post-famine cohort differences in the DID estimates by examining the variations in subgroups with different types of infectious diseases (trachoma, poliomyelitis, tuberculosis, maternal infections and other infectious diseases). ConclusionNo significant impact of prenatal exposure to the Chinese famine was observed on the risk of infectious diseases in adulthood. Famine survivors may be “cured” by the famine and were resilient to adverse environments in their life course because selective mortality may weaken the association between adverse prenatal exposure and later health.

Analysis of State Medicaid Expansion and Access to Timely Prenatal Care Among Women Who Were Immigrant vs US Born

Disparities exist in access to timely prenatal care between immigrant women and US-born women. Exclusions from Medicaid eligibility based on immigration status may exacerbate disparities. To examine changes in timely prenatal care by nativity after Medicaid expansion. A cross-sectional difference-in-differences (DID) and triple-difference analysis of 22 042 624 singleton births from January 1, 2011, to December 31, 2019, in 31 states was conducted using US natality data. Data analysis was performed from February 1, 2021, to August 24, 2022. Within 16 states that expanded Medicaid in 2014, the rate of timely prenatal care by nativity in years after expansion was compared with the rate in the years before expansion. Similar comparisons were conducted in 15 states that did not expand Medicaid and tested across expansion vs nonexpansion states. Timely prenatal care was categorized as prenatal care initiated in the first trimester. Individual-level covariates included age, parity, race and ethnicity, and educational level. State-level time-varying covariates included unemployment, poverty, and Immigrant Climate Index. A total of 5 390 814 women preexpansion and 6 544 992 women postexpansion were included. At baseline in expansion states, among immigrant women, 413 479 (27.3%) were Asian, 110 829 (7.3%) were Black, 752 176 (49.6%) were Hispanic, and 238 746 (15.8%) were White. Among US-born women, 96 807 (2.5%) were Asian, 470 128 (12.1%) were Black, 699 776 (18.1%) were Hispanic, and 2 608 873 (67.3%) were White. Prenatal care was timely in 75.9% of immigrant women vs 79.9% of those who were US born in expansion states at baseline. After Medicaid expansion, the immigrant vs US-born disparity in timely prenatal care was similar to the preexpansion level (DID, -0.91; 95% CI, -1.91 to 0.09). Stratifying by race and ethnicity showed an increase in the Asian vs White disparity after expansion, with 1.53 per 100 fewer immigrant women than those who were US born accessing timely prenatal care (95% CI, -2.31 to -0.75), and in the Hispanic vs White disparity (DID, -1.18 per 100; 95% CI, -2.07 to -0.30). These differences were more pronounced among women with a high school education or less (DID for Asian women, -2.98; 95% CI, -4.45 to -1.51; DID for Hispanic women, -1.47; 95% CI, -2.48 to -0.46). Compared with nonexpansion states, differences in DID estimates were found among Hispanic women with a high school education or less (triple-difference, -1.86 per 100 additional women in expansion states who would not receive timely prenatal care; 95% CI, -3.31 to -0.42). The findings of this study suggest that exclusions from Medicaid eligibility based on immigration status may be associated with increased health care disparities among some immigrant groups. This finding has relevance to current policy debates regarding Medicaid coverage during and outside of pregnancy.