A binary variable indicated whether the period was pre- or post-policy

While the use of ACHS to capture the outcomes of all arrests in the state is a strength of this study, ACHS also faces the quality challenges typical of large administrative datasets, as CA DOJ must rely upon consistent and timely reporting from 58 counties. Though courts and law enforcement agencies are mandated to report within 30 days of final case dispositions and the CA DOJ’s policy is to update the data system within 90 days of receipt, a substantial portion of arrests did not contain dispositions. We assumed that these arrests without dispositions were not prosecuted for the primary analysis. However, if cases with no dispositions in fact include some felony convictions, and felony conviction missingness is associated with county, it could contribute to some of the geographic variation in convictions. The analysis of change in variation across time could be biased if felony conviction missingness differed within counties in the year pre- vs. post-Prop 47. There are several pieces of evidence that provide some reassurance. First, missing dispositions were more likely in the post period, which we would expect to occur if missing dispositions were indicative of no conviction, since the classification of drug possession offenses was reduced. Second, cases with missing dispositions were less severe in terms of concurrent offenses,cannabis grow lights which would correspond with lower likelihood of felony conviction. Third, the sensitivity analysis assuming that cases with missing dispositions had resulted in felony convictions did not alter findings. The impact of the study design on the potential for bias should be considered.

By comparing events just within the year before and after Prop 47, we attempted to limit the effect of time trends in felony convictions, though some reduction in felony convictions could be attributed to a pre-existing trend towards leniency for drug possession. That said, the large and immediate reduction in felony convictions across nearly all counties is unlikely to have occurred in the absence of the policy change. Leading up to and following the passage of Prop 47, there was substantial debate about potential unintended consequences of reducing criminal penalties for all drugs in California. For example, reductions in incarceration for drug offenses may lead to growth in the number of people with SUDs and other behavioral health conditions in the community. Increases in the needs of community members may not be adequately met by available community-based services . One criticism of Proposition 47 was the possibility that it would reduce treatment previously accessed in the context of drug diversion programs . Drug diversion programs offer eligible individuals with drug-related arrests the option to attend drug treatment in lieu of incarceration, and to have charges dismissed if treatment is completed . Some research suggests diversion programs have had success in reducing substance use and recidivism, and legally mandated treatment can provide an important opportunity for linkage to care . For those with misdemeanor offenses the opportunity to divert a felony charge and/or incarceration is incentive to enroll into drug diversion programs. Prop 47 may remove this incentive by altering the alternative choice from a felony to a misdemeanor, with the maximum penalty likely to be a short jail sentence and misdemeanor probation. Those arrested post-Prop 47 may prefer this option to drug diversion programs that dictate lengthy terms of treatment with strict conditions for completion, and the possibility of more severe penalties if conditions are not met. Although fewer may access drug diversion options, there are also reasons to hypothesize reducing criminal penalties may minimize harms to substance users.

Some evidence suggests that incarceration may exacerbate substance use . Jails and prisons rarely offer harm reduction strategies such as methadone maintenance; forced withdrawal lowers post-release treatment engagement by seven-fold , and may cause increased risk of injecting drugs after incarceration . The break or change in substance use due to time incarcerated reduces one’s tolerance, contributing to the high risk of overdose in the first two weeks post-release . Taken together, it is uncertain how reducing drug possession penalties would affect rates of overdose. Media reports in 2015 and 2017 suggested Prop 47 led to upticks in drug-related ED visits among individuals who would have been treated or incarcerated prior to the policy change , though a statewide study has not been conducted to date. As other states increasingly enact similar changes in drug laws, it is critical to better understand the health implications. This study assesses whether there was a change in SUD-related hospital visits statewide, following the reduction in criminal penalties for drug possession in California. We also evaluate whether county-level arrest rate changes associated with Prop 47 predict county-level changes in SUD-related hospital visits. To evaluate whether changes in SUD-related hospital visits were attributable to Prop 47, we assessed whether county-level changes in drug arrest rates were associated with county-level changes in SUDrelated hospital visit rates. The measure of change for both variables was defined as the difference in the observed and expected rates during the ten months post-policy period, for each county and gender-age group. To estimate expected rates, we modeled monthly county rates with cluster robust standard errors using a Poisson specification and county fixed effects, stratifying models by gender-age group. Calendar month dummies controlled for seasonal trends, and a continuous linear term for months controlled for secular trends.To project post-policy rates based on pre-policy trends, we created duplicate observations but set the post-policy binary variable to zero, such that model predictions were based on how trends were expected to proceed in the absence of the policy.

We then calculated the model-predicted rate for the 10-month post-policy period, and subtracted it from the observed rate. Linear regression was used to model the effect of a reduction of 100 drug arrests per 100,000 on the SUD-related hospital visit rate in each county or county grouping during the 10-month post-policy period. All analyses were conducted in Stata version 15.Counties with small populations were grouped by region, cannabis grow tent replicating the groupings created by the Office of Statewide Health Planning and Development . Groups included North Central , Foothills , Central Coast , North Coast , North , North East , and East Central . A total of 285,575 SUD-related hospital visits occurred in California from October 5, 2011 – September 4, 2015 among individuals ages 15-64, approximately one third of which resulted in an inpatient admission . Rates were higher among males, particularly White males ages 15-24, than other groups. Though drug types were not mutually exclusive , opioids were coded in the largest proportion of visits , followed by amphetamines . Visits coded with opioids were also most likely to include multiple drugs , while no visits for amphetamines had multiple drugs listed. Changes in statewide SUD-related hospital visits could be attributable to a variety of factors unrelated to Prop 47. Since the immediate effect of Prop 47 was to reduce drug arrest rates, if changes in hospital visits were related to Prop 47, counties with greater reductions in drug arrests would be expected to experience greater changes in hospital visits. Table 4.2 displays the median observed vs. expected county drug arrest and SUD-related hospital visit rates per 100,000 in the 10 months post-policy, by gender-age group. We find that median county drug arrest rates declined in every group, with the greatest declines among males ages 15-24 . Median increases in SUD-related hospital visit rates were also greatest in this population , which reflects findings from the statewide analysis. However, results from the linear regression of county changes in SUD-related hospital visits on county changes in drug arrest rates indicates associations were non-significant for every age and gender group, including males ages 15-24 . Though the confidence interval included zero, the direction of the effect was the reverse of what was expected: among males ages 15-24, a decline of 100 per 100,000 drug arrests was associated with a decline of 8 per 100,000 SUD-related hospital visits .This study of the effects of Prop 47 on SUD-related hospital visits in California suggests that a reduction in drug arrests was not associated with an increase in hospital visits.

Evidence shows that increasing the severity of sanctions for drug offenses historically did not affect use , though research on the health effects of easing drug laws has been generally limited to marijuana. This body of literature suggests that medical marijuana legalization had no effect on marijuana initiation or DSM-IV dependence/abuse among users , and reduced opioid-related hospital admissions with no impact on marijuana related hospitalizations . Studies of the impacts of legalizing marijuana for recreational use show greater divergence and potentially negative health effects, with increases in emergency department visits, hospitalizations, and regional poison control center calls linked to marijuana exposure , and mixed findings regarding effects on adolescent use and perceived harmfulness across states . Legalization, however, is distinct from reducing criminal penalties, with different mechanisms through which use could be altered. For example, though legalization may have a similar effect as reducing criminal penalties on minimizing arrests and incarceration associated with the drug, it also increases access to the drug. Legalization may garner media attention that increases public awareness of the reform, in turn potentially reducing perceptions of risks associated with the drug or generating drug use tourism . Reclassifying drug offenses, in contrast, may go relatively under the public radar, which could correspond with a lesser effect on public perceptions or use patterns. Regardless of public awareness of the reform, there is clear evidence that arrests and incarceration dropped as a result of Prop 47 . This study indicates reducing drug arrests did not alter SUD-related hospital visits on a population level. Findings would be bolstered by further research that examines the effect of shorter sentences in addition to reduced arrest rates, and that investigates changes in individual trajectories through the criminal justice system, access to treatment, and outcomes. While individual-level studies would offer a stronger assessment of the association between the exposure and outcome of interest, causal inference in such studies would be challenged by the likelihood that police and prosecutors shifted focus to individuals with more severe criminal histories post-policy. Such a shift would itself be worth exploring; if the individuals no longer getting arrested are those with a lower risk of hospital visits, then it may help to explain why the drop in arrests had little effect on hospital visit rates. In any case, it is imperative to gain insight into whether and how individuals at greatest risk of negative health outcomes from substance use are being engaged in services within the new policy environment, including best practices for counties’ use of Prop 47 savings to increase treatment access. This study was limited by a brief post-policy time period of 10 months; continued monitoring will help to shed light on the full impact of the policy, particularly after Prop 47 county grants are invested in treatment. Though reductions in drug arrests were not associated with a rise in hospital visits, it is plausible that other measures of change in criminal justice involvement attributable to Prop 47, such as time incarcerated, may have different effects that could be explored. Finally, hospital visits do not provide a comprehensive picture of the effects of drug law reforms on substance use and corresponding outcomes. Further research might examine reported use, SUD treatment access, criminal justice involvement among substance users, and measures of well-being such as employment, housing, mental health, and familial bonds. Despite an increase in female-only studies on HIV associated neurocognitive impairment , few compare HIV-positive women and men. Women are often, but not always, more likely to demonstrate NCI. This female-specific vulnerability may reflect a greater prevalence of psychosocial factors in women that have biological effects on the brain. These factors or ‘syndemics’ may additively or synergistically harm cognitive/brain health and lower cognitive reserve. Biopsychosocial syndemics are associated with HIV acquisition, antiretroviral adherence and may also contribute to NCI. In the Women’s Interagency HIV Study , reading level, years of education and income were more strongly associated with cognition than HIV-serostatus, and stress/anxiety increased risk of NCI among HIV-positive, but not HIV-negative, women. Although biopsychosocial factors may increase HIV-associated NCI, studies have not examined the role of biopsychosocial syndemics in sex differences in risks for HIV-associated NCI.