Utilization of emergency department resources are 50% to 100% higher for patients with SUD compared with patients without SUD.In addition to acute medical emergencies, ED use may be indicative of poor health, unmet service need, or inappropriate use of health care.To date, studies have found most SUD-related ED visits are associated with alcohol,and frequently document ED-based treatments have focused on alcohol to the exclusion of other drugs.Yet, ED visits associated with the misuse of opioids and marijuana are common, and considerable SUD-related ED visits involve concurrent or other drug use.In addition, alcohol and opioid use disorders are among the most severe SUD diagnoses in terms of their negative impact on health, and evidence continues to emerge about the adverse health effects associated with marijuana use disorder.Thus, the study of ED trends among patients with alcohol, marijuana, and opioid use disorders is important.High rates of SUD-related clinical emergencies and associated ED visits are a persistent barrier to improving health outcomes in this population.Thus, a study that seeks to identify how patients with alcohol, marijuana, and opioid use disorders use ED resources is important, to potentially inform more specific ED-based treatment efforts . This study examined ED trends across patients with alcohol, marijuana, and opioid use disorders, and controls, over time in a large integrated health care system in which all patients have insurance coverage to access health care. Using electronic health record data, we aimed to determine the odds of having an ED visit each year from 2010 to 2014 for patients with alcohol, marijuana, and opioid use disorders relative to controls without these conditions; evaluate differences in ED use between controls and those with alcohol, cannabis grow set up, and opioid use disorders over 5 years; and explore sub-samples for which patients with SUD may have a greater impact on ED resources.We used secondary EHR data for this database-only study.
These data were used to identify all health plan members who were aged 18 or older, who had a visit to a KPNC facility in 2010, and had a recorded ICD-9 diagnosis of alcohol, marijuana, or opioid abuse or dependence in 2010. The first mention for each ICD-9 diagnosis of alcohol, marijuana, or opioid use disorder recorded from January 1, 2010, to December 31, 2010, were included; patients in the sample could have multiple diagnoses . We also included all current or existing SUD diagnosis that were additionally documented for patients with alcohol, marijuana, or opioid use disorder during health plan visits in 2010 . Within KPNC, SUD and other behavioral health diagnoses can be assigned to patients in any clinic setting, e.g., primary care or any specialty care clinic.All diagnoses are captured through ICD-9 codes. EHR data were used to identify control patients who did not have current or existing SUDs or other behavioral health diagnoses. Control patients were selected for all unique patients with alcohol, marijuana, and opioid use disorders and matched one-to-one on gender, age, and medical home facility. This accounted for differences in services, types of behavioral health conditions, or unobservable differences by geographic location. To control for varying lengths of membership, participants were required to be KPNC members for at least 80% of the study . The final analytical sample consisted of 35,148 patients: 12,411 with alcohol use disorder, 2752 with marijuana use disorder, 2411 with opioid use disorder, and 17,574 controls. Institutional review board approval was obtained from the Kaiser Foundation Research Institute.Overall, the sample was 35.5% women, 60.0% white, 16.1% Hispanic, 11.0% Asian, 8.6% black, and 4.0% other race/ethnicity. Patients were 37 years old on average. Differences in the characteristics among patients with alcohol, marijuana, and opioid use disorders and the controls are reported in Table 1. Compared with controls, more patients with alcohol, marijuana, or opioid use disorder were white or black; more controls were Asian, Hispanic, or had a race/ethnicity categorized as “other” compared with those with alcohol, marijuana, and opioid use disorder with few exceptions. In addition, compared with controls, patients with alcohol, marijuana, and opioid use disorders had greater medical comorbidities , and co-occurring mental health and substance use conditions were common .
Alcohol, marijuana, and opioids frequently take center stage in public policy and debate as concerns remain focused around opioid misuse and overdose,ongoing drinking problems,and liberalization of marijuana use policies.Persons who excessively use these substances face the risk of developing an associated SUD,which can have considerable implications for patient health and health systems,in part by contributing to high use of ED services.Thus, we examined how patients with alcohol, marijuana, and opioid use disorders, and controls, used ED resources over time in a large health care system. Similar to studies conducted in the general population and other health systems,alcohol use disorder was diagnosed the most frequently, followed by marijuana use disorder, and opioid use disorder, and the rates of cooccurring medical, psychiatric, and SUD were substantial in each. Because these conditions worsen prognosis, lead to high morbidity,and can contribute to inappropriate service use,it is not surprising we found that patients with these disorders consistently had greater likelihood of ED use relative to controls. ED visits were the highest among patients with opioid use disorder, followed by those with marijuana and alcohol use disorders, which is contrary to prior work that has documented most SUD related ED visits are associated with alcohol use disorder.This difference could reflect the effects of changing marijuana use disorder patterns and an overall high morbidity among patients with opioid disorder, which may have large effects on health system resources.Most ED-based treatments focus on alcohol to the exclusion of other drugs,and since our data suggest that ED visits are also frequent among patients with marijuana and opioid use disorders, these patients may be at risk for having unmet or unidentified treatment needs. Consequently, building on ED based treatments for patients with alcohol use disorder,it will be important for future studies to extend these treatments to patients with opioid and marijuana use disorders, to reduce medical emergencies and improve patient health in this population. Patients with opioid use disorder constituted a modest proportion of the sample, and these patients consistently had high odds of ED use. Similar to this, previous studies report that patients with opioid use disorder are over represented in ED settings.This could be due to the individual or combined effects of complex medical conditions, injury, or overdose,which have large impact on the burden of disease and are some of the more persistent barriers to improving overall health outcomes among patients with opioid use disorder.
Consequently, ED settings offer important opportunities to identify patients with opioid use disorder and initiate treatment. Recent evidence suggests that ED-initiated buprenorphine increases subsequent engagement in addiction treatment and reduces illicit opioid use.Devoting more health resources to initiating evidence-based ED-based treatments for patients with opioid use disorder in health systems, including ED-initiated buprenorphine and referral to SUD treatment,may be a step toward improving health outcomes and reducing high SUD related ED visits among patients with opioid use disorder. Over time, all patients had fewer ED visits, and a greater decrease in ED use was observed for patients with SUDs compared with controls, although those with SUDs continued to have more ED visits. These ED utilization patters are consistent with general population studies, which show decreasing ED visits involving alcohol and opioid use disorders.At the same time, our ED utilization patterns regarding marijuana use disorder are inconsistent with national data, which suggest increasing ED visits involving marijuana-related problems.This national increase could be due to the combined effects of increasing marijuana potency, liberalizing views of the drug, and increasing trends toward its legalization.Notably, however, we found a decrease in ED use over time across patients with marijuana use disorder as well as those with alcohol and opioid use disorders, which may suggest that some patients’ health status improves more quickly. Another possibility is that the observed decrease in ED use may be specific to those who receive care within integrate health systems in which specialty services are provided internally. For example,outdoor cannabis grow prior studies conducted within KPNC found that patients with SUD who had ongoing primary care and addiction treatment were less likely to have subsequent ED visits.It will be important for future studies in other systems to investigate the potential impact of specialty and primary care on reducing subsequent acute services across those with alcohol, marijuana, and opioid use disorders. Our results confirm the work of prior studies showing that patients with alcohol and opioid use disorders, and to a lesser degree patients with marijuana use disorder, have frequent and increasing ED visits over time associated with poor health or complex medical conditions.Since our medical comorbidity measure combined acute and chronic conditions, it will be important for future work to identify which individual medical conditions contribute most strongly to ED admission. Other characteristics that were not measured may also influence ED use rates in patients with SUD, and understanding these factors may further help improve service planning efforts and ED-based treatments for this population. In addition, comorbid conditions were common among patients with SUD, and these individuals may have ED visits that require a range of medical treatments, psychiatric symptom stabilization, or detoxification from alcohol or drugs. Limitations should be noted. Our use of provider-assigned diagnoses restricted the sample to patients with at least 1 of the 3 most common SUD diagnoses in 2010 . As with other studies that have used claims-based data,our study captures patients with SUD through ICD-9 codes noted in health plan visits during the study period. This methodology is vulnerable to diagnostic underestimation.Therefore, the SUD prevalence data in our study may underestimate the general ED patient population prevalence.
Although not available for this study, future database studies could examine if the inclusion of pharmacy-based prescription data to ICD-9 diagnosis improves prevalence estimates. Another potential limitation with the methods we used to select our SUD sample is that we required a single mention of an ICD-9 code for SUD during the study period to link the patient with that diagnosis. Although the single mention methodology is well established,it could result in an overestimation of the true diagnostic rates if diagnoses only mentioned one time in the EHR are more likely to be inaccurate. Patients were insured members of an integrated health system, and thus our results may not be generalizable to uninsured populations or other types of health systems. Our findings of SUD-related ED trends are somewhat inconsistent with prior work,which suggests a need for replication. All patients were required to have a health system visit in 2010 to enter the study, but they were not required to have a health system visit to remain in the study. These criteria may explain the steep decline in ED visits between 2010 and 2011 and subsequent leveling of ED use. We cannot identify the reason for why patients had an ED visit , which will be an important focus of future work. ED utilization that KPNC did not pay for is not captured, although we capture external, paid-for ED utilization through claims. Consequently, ED use may be higher than we report. Low base rates of SUDs other than alcohol, marijuana, and opioid use disorders precluded our ability to examine the effect of these conditions on ED visits.The empirical literature documents statistically and clinically significant relations between HIV/AIDS and anxiety and depressive symptoms and disorders . Rates of anxiety disorders among HIV? individuals have been estimated as high as 43 % . Likewise, depressive symptoms and disorders commonly co-occur with HIV/AIDS, with some studies finding over a 50 % base rate of clinical depression among adults with HIV/AIDS . Although the underlying directionality between anxiety and depressive symptoms and disorders and HIV/AIDS is presently unclear, research has nonetheless found that these negative emotional states tend to contribute to non-adherence to HIV medications , lesser quality of life , greater health-care utilization , and greater risky sexual behaviors . Scholars have begun to focus greater energy on identifying the explanatory processes that may underlie such anxiety/depression-HIV/AIDS associations. The most well developed aspect of this literature has been focused on coping with the HIV/AIDS illness and other life stressors . Yet, there has been little investigation of other cognitive-affective factors related to these negative emotional states.