Prevalence and trends in cannabis use disorder and cannabis poisoning among Medicaid enrollees: a multistate analysis, 2011–2022
Department of Health Management and Policy, College of Public Health, University of Kentucky, Lexington, KY 40536, United States
Department of Pharmacy Practice and Science, College of Pharmacy, University of Kentucky, Lexington, KY 40536, United States
Corresponding author: Department of Health Management and Policy, College of Public Health, University of Kentucky, Lexington, KY 40536, United States. Email: jayani.jayawardhana@uky.eduAbstract
Introduction
Cannabis use in the United States is increasing. However, the prevalence and trends in cannabis use disorder (CUD) and cannabis poisoning among Medicaid enrollees, a vulnerable population, are not evident.
Methods
Using the Merative MarketScan Multistate Medicaid Claims and Encounters Database from 2011–2022, we examined adjusted prevalence and trends in CUD and cannabis poisoning among Medicaid enrollees and by age, sex, and insurance type (managed care [MC]/fee-for-service [FFS]).
Results
During the 2011 quarter (Q) 1–2022 Q4, the CUD rate increased from 336.54 to 548.96 per 100 000 enrollees per quarter—a 1.63-fold increase; the cannabis poisoning rate increased from 1.45 to 7.04 per 100 000 enrollees per quarter—a 4.86-fold increase. CUD rates were highest among those aged 18–34 years, while cannabis poisoning rates among those aged 0–17 years surpassed the rate of those aged 18–34 years by 2020 Q3. CUD and cannabis poisoning rates increased among both males and females and among those with FFS and MC, although females and MC enrollees experienced higher increases than males and FFS enrollees, respectively.
Conclusion
CUD and cannabis poisoning rates among Medicaid enrollees increased significantly during 2011–2022, especially among older adults, females, and MC enrollees. Targeted education campaigns on safe use and storage of cannabis may help reduce increasing trends in CUD and cannabis poisonings.
Introduction
Cannabis is classified as a schedule I drug under the Controlled Substances Act of 1970 and is federally illegal in the United States.1 However, many states have legalized cannabis both for medical and recreational purposes, while others have decriminalized cannabis use.2 As a result, cannabis use has been increasing in the United States,3,4 with more potent cannabis products becoming available in the market5-7 and more people perceiving cannabis as harmless.8,9 However, with the increased availability of cannabis through legalization, its use has been linked to adverse health outcomes, including increased rates of cannabis use disorder (CUD) and cannabis poisoning in different populations.4,10-14
Cannabis use disorder is a psychiatric disorder described as continued use of cannabis despite significant impairment to one's health and/or social circumstances.15,16 It is reported that approximately 30% of cannabis users are likely to develop CUD, and the risk of developing CUD is higher among frequent cannabis users and those starting cannabis use in adolescence.16,17 Cannabis use disorder has negative effects on mental, physical, and social health, leading to adverse health conditions, such as psychosis, anxiety, sleep disorders, and cardiovascular problems.18 Cannabis use disorder can also affect cognitive function and increase the risk of accidents.18,19 A higher prevalence of CUD in veteran and commercially insured populations has been reported in states where medical and recreational cannabis is legal compared with states without such legalization.10,14
Consuming excessive amounts of cannabis or highly potent cannabis at one time can result in cannabis poisoning.20 Edible cannabis has a higher risk of poisoning than smoked cannabis due to its delayed and longer-lasting effects and the difficulty in measuring or knowing its potency accurately.21 Children are more susceptible to cannabis poisoning than adults since children are more sensitive to cannabis effects and because cannabis can be mistaken for candy or food and unknowingly eaten by children.18,20,21 Although cannabis poisoning may not be fatal, it can result in unpleasant and terrifying symptoms, including panic attacks, palpitations, paranoia, hallucinations, nausea, dizziness, vomiting, dry mouth, and impaired coordination, necessitating emergency care.22-24 The effects of cannabis poisoning may vary across individuals and depend on the potency of the product consumed, the method of consumption, the individual's tolerance levels, and their other health risks.22,23,25
The Medicaid population is a vulnerable, underserved population with a higher prevalence of substance use disorders (SUDs), including CUD, than the general US population, and is at a higher risk of developing CUD.26 Medicaid enrollees also experience limited access to preventive or specialty care services compared with other populations.27 Therefore, understanding patterns of their CUD and cannabis poisoning diagnoses can help inform policymakers and clinicians on the extent of the conditions, and help guide clinical and policy interventions. In 2019, approximately 2% of Medicaid enrollees reported being diagnosed with CUD, while 3.3% of Medicaid enrollees were diagnosed with opioid use disorder. However, more recent data are lacking.26 Very few studies have examined cannabis use among Medicaid enrollees,28-30 and none have examined the prevalence and trends in CUD and cannabis poisoning in this population. It is also not evident if CUD and cannabis poisoning rates in this population differ by age, sex, or payer type (managed care [MC]/fee-for-service [FFS]) subgroups. The literature to date has mostly focused on privately insured individuals or the general population,10,12,14,31 leaving a gap in understanding unique experiences by Medicaid enrollees, a low-income, high-risk population. To fill this gap in the literature, this study examines the prevalence and trends in the rates of CUD and cannabis poisoning in a sample of Medicaid enrollees from 2011 through 2022 overall and by their subgroup categories (age, sex, and payer type) using MarketScan Medicaid claims data. Findings of this study could inform policymakers on coverage policies related to SUD treatment and inform Medicaid programs on where to direct behavioral health services, provider training, or public health interventions.
Data and methods
Data
This study used Merative MarketScan Multi-State Medicaid data from January 1, 2011, to December 31, 2022. This database includes pooled health care encounters of Medicaid enrollees living in numerous geographically dispersed states covered under FFS and MC plans and includes both inpatient and outpatient records. Medicaid enrollees with CUD were identified using International Classification of Diseases, Ninth Revision–Clinical Modification (ICD-9-CM) codes 305.2X and 304.3X and International Classification of Diseases, Tenth Revision–Clinical Modification (ICD-10-CM) codes F12.1 and F12.2, which indicate cannabis abuse and dependence. ICD-9-CM codes 969.6, E854.1, and E939.6 and ICD-10-CM codes T40.7, indicating accidental and intentional cannabis poisoning and adverse effects of cannabis, were utilized to identify enrollees with cannabis poisoning incidents. These diagnoses were recorded whenever an inpatient or outpatient encounter took place. The Merative MarketScan Multi-State Medicaid data do not contain state identifier information. The study sample included 30 408 341 Medicaid enrollees during the study period. The study was approved by the Institutional Review Board at the authors’ institute.
Measures
The primary outcome variables were the adjusted rates of CUD and nonfatal cannabis poisoning (hereafter referred to as cannabis poisoning) diagnoses per 100 000 enrollees per quarter (Q). Since the demographic composition of the study sample could change over time, we adjusted for the composition of age, sex, and insurance type using the total number of enrollees in the initial period (2011 Q1) and their demographic composition as the reference group and standardized the rates of CUD and cannabis poisoning for the rest of the study period. The crude CUD rate per 100 000 enrollees per quarter was calculated as the number of enrollees with a CUD diagnosis in a given quarter divided by the total number of enrollees in that quarter and multiplying the quotient by 100 000. Then, these crude CUD rates were multiplied by the standardized weights that were calculated using the 2011 Q1 demographic composition of the study sample to yield the weighted CUD rate for a given year-quarter. A similar approach was followed to calculate the adjusted cannabis poisoning rate per 100 000 enrollees per quarter. An individual was counted once in every quarter if they were diagnosed with CUD or cannabis poisoning in that quarter. The secondary outcome variables were the adjusted rates of CUD and cannabis poisoning diagnoses per 100 000 enrollees by sex (male/female), age groups (0–17, 18–34, 35–64, ≥65 years), and insurance type (MC/FFS) per quarter.
Statistical analysis
The quarterly adjusted rates of CUD and cannabis poisoning per 100 000 enrollees were calculated as described above in the Measures section. The trends in the adjusted rates of CUD and cannabis poisoning were examined in the overall study sample and by their subgroup characteristics (age, sex, MC/FFS) using graphical analyses. We utilized the Prais-Winsten regression approach to assess statistical significance of the trends over time, adjusting for first-order autocorrelation (AR[1]). This approach provides more reliable estimates compared with ordinary least-squares regression, which assumes independent errors.
Results
Appendix Table S1 presents the summary statistics of the study sample. Approximately 55.54% of the study sample was female. On average, in a given quarter, approximately 50.75% (SE = 5.42%) of enrollees had managed care, although it is important to note that Medicaid enrollees switched between MC and FFS during the study period. Since Medicaid enrollees’ age changed over the study period, we calculated the average percentage of age groups likely to be present in each year-quarter. On average, approximately 48.61% (SE = 4.87%) were in the 0–17-year age group, 21.25% (SE = 1.68%) were in the 18–34-years age group, and 22.02% (SE = 3.17%) were in the 35–64-year age group.
Figure 1 presents adjusted trends of overall rates of CUD and cannabis poisoning diagnoses during 2011–2022. The adjusted rate of CUD among Medicaid enrollees increased from 336.54 to 548.96 patients per 100 000 enrollees per quarter, resulting in a 1.63-fold increase from 2011 Q1 to 2022 Q4. The Prais-Winsten regression slope coefficient of the CUD rate for year-quarter was statistically significant (β = 5.40; P < .001), indicating a positive trend (Appendix Table S2). Similarly, the adjusted cannabis poisoning rate among Medicaid enrollees increased from 1.45 to 7.04 patients per 100 000 enrollees per quarter, resulting in a 4.86-fold increase during the same period. The estimated slope coefficient of cannabis poisoning rate for year-quarter was statistically significant (β = 0.13; P < .001), indicating a positive trend (Appendix Table S2).
Figure 2 presents adjusted trends of CUD and cannabis poisoning among Medicaid enrollees by age groups 0–17, 18–34, 35–64, and 65-plus years during the 2011–2022 period. The adjusted rate of CUD was highest among those aged 18–34 years throughout the study period. The adjusted rate of CUD increased from 66.40 to 95.18, from 171.52 to 260.77, from 97.51 to 182.26, and from 1.12 to 10.76 patients per 100 000 enrollees per quarter among those aged 0–17, 18–34, 35–64, and 65-plus years, respectively, resulting in a 1.43-fold, a 1.52-fold, a 1.87-fold, and a 9.61-fold increase in each group, respectively. The estimated slope coefficients of CUD rate for year-quarter across all age groups except for those aged 0–17 years were statistically significant (P < .001), indicating positive trends (Appendix Table S2). The adjusted rate of cannabis poisoning increased from 0.47 to 4.57, from 0.57 to 1.54, from 0.36 to 0.80, and from 0.04 to 0.13 patients per 100 000 enrollees per quarter among those aged 0–17, 18–34, 35–64, and 65-plus years, respectively, resulting in a 9.72-fold, a 2.70-fold, a 2.22-fold, and a 3.25-fold increase in each group, respectively. The regression estimated slope coefficients of cannabis poisoning rate for year-quarter across all age groups were statistically significant (P < .001), indicating positive trends (Appendix Table S2).
Figure 3 presents adjusted trends of CUD and cannabis poisoning among Medicaid enrollees by sex during the 2011–2022 period. The adjusted rates of CUD and cannabis poisoning increased throughout the study period among both males and females. Although males experienced higher rates of CUD and cannabis poisoning than females at the beginning of the study period (2011 Q1), females experienced higher rates than males by the end of the study period (2022 Q4). The adjusted rate of CUD among males increased from 193.79 to 243.51 patients per 100 000 enrollees per quarter, resulting in a 1.26-fold increase from 2011 Q1 to 2022 Q4. The adjusted rate of CUD among females increased from 142.75 to 305.45 patients per 100 000 enrollees per quarter, resulting in a 2.14-fold increase during the same period. The adjusted cannabis poisoning rates increased from 0.84 to 3.19 (males) and from 0.61 to 3.85 (females) patients per 100 000 enrollees per quarter, resulting in a 3.80-fold increase among males and a 6.31-fold increase among females, during 2011 Q1–2022 Q4. The Prais-Winsten regression estimated coefficients of the slope for year-quarter for both outcomes among both males and females were statistically significant (P < .001), indicating positive trends (Appendix Table S2).
Figure 4 presents adjusted trends of CUD and cannabis poisoning among the Medicaid enrollees by insurance plan types (MC/FFS) during the 2011–2022 period. The adjusted rates of CUD and cannabis poisoning increased throughout the study period across those with MC and FFS. The adjusted rate of CUD among those with MC increased from 156.16 to 265.25 patients per 100 000 enrollees per quarter, resulting in a 1.70-fold increase during the study period. The adjusted rate of CUD among those with FFS plans increased from 180.38 to 283.71 patients per 100 000 enrollees per quarter during the same period, resulting in a 1.57-fold increase. The estimated slope coefficients of CUD rate for year-quarter among those with MC (β = 2.75) and FFS (β = 2.05) were statistically significant (P < .001), indicating positive trends (Appendix Table S2). The adjusted rates of cannabis poisoning increased from 0.53 to 3.51 and from 0.92 to 3.53 patients per 100 000 enrollees per quarter, resulting in a 6.62-fold increase among those with MC and a 3.84-fold increase among those with FFS plans. The estimated slope coefficients of cannabis poisoning rate for year-quarter among those with MC (β = 0.07) and FFS (β = 0.06) were statistically significant (P < .001), indicating positive trends (Appendix Table S2).
Discussion
In this study, we examined adjusted trends of the number of patients with CUD and cannabis poisoning diagnoses among Medicaid enrollees during the 2011 through 2022 period. The results show that the quarterly adjusted rate of patients with CUD diagnoses per 100 000 enrollees increased by 1.63-fold and the quarterly adjusted rate of patients with cannabis poisoning diagnoses per 100 000 enrollees increased by 4.86-fold during the study period. Although cannabis is federally illegal in the United States, the study period captures a changing cannabis policy landscape, where many states have legalized cannabis for medical and recreational purposes while others have decriminalized cannabis use.2 Such changes at the state level are likely to result in increased social acceptance of cannabis as harmless,8,9,32 leading to increased cannabis use,4,33 which, in turn, is likely to increase the prevalence of CUD and cannabis poisonings.10,14,31
Our findings also show that those aged 18–34 years experienced the highest adjusted prevalence rates of CUD throughout the study period compared with other age groups. Increased cannabis use in this age group than in other age groups34 and multimodal use of cannabis among young adults35-37 are likely contributing to higher CUD rates among those aged 18–34 years. Interestingly, those aged 65 years and older experienced the greatest increase in adjusted CUD rate during the study period, a 9.61-fold increase. Recent studies have shown increased use of cannabis among older adults,38,39 which may lead to increased CUD diagnoses. While those aged 18–34 years experienced higher adjusted cannabis poisoning rates than other age groups up until 2020 Q2, the adjusted cannabis poisoning rate among those aged 0–17 years surpassed other age groups starting 2020 Q3 and continued to increase for the rest of the study period. Additionally, the greatest increase in adjusted cannabis poisoning prevalence rate was experienced by the 0–17-year age group, a 9.72-fold increase, while older adults (≥65 years) experienced the second highest increase in adjusted cannabis poisoning prevalence rate, a 3.25-fold increase, during the study period. The increase in cannabis poisoning prevalence rate among the 0–17-year age group is particularly concerning since children are more vulnerable to cannabis poisoning effects and may experience more severe symptoms than adults.20,21 Increased access to cannabis due to legalization,40,41 increased exposures to edible cannabis,41,42 and accidental ingestion of cannabis due to unsafe storage or product appearance20,21,24 are likely contributing factors for higher cannabis poisoning prevalence among those aged 0–17 years. Our results, however, are consistent with previous findings that show increased reporting of pediatric poisoning incidences in regions where edible cannabis sales are legal,41 and increased reporting of cannabis exposures among those aged younger than 21 years to the US National Poison Data System where recreational cannabis commercialization is legal.43 The finding of increasing cannabis poisoning rates among older adults is also disturbing since they are more susceptible to cannabis effects due to their worsening physical health; higher use of prescription medications, which may lead to interactions with cannabis; and inability to accurately recognize the potency of cannabis products.44,45 Our findings are in line with previous findings that show increased use of cannabis among older adults and increased cannabis-related emergency department visits among older adults (≥65 years), especially with the availability of edible cannabis.44-46
While both males and females experienced increasing adjusted prevalence rates of CUD and cannabis poisoning throughout the study period, females experienced the greatest increases in both CUD and cannabis poisoning prevalence rates, with 2.14-fold and 6.31-fold increases during the study period, respectively. Consistent with our results, a previous study showed higher relative increases in CUD and cannabis poisoning among females than among males in states where medical cannabis is legal.14 A recent survey also reported higher use of cannabis among young women aged 19–30 years than among men of similar age.47 Although, in general, men are more likely to use cannabis than women, the sex gap in the prevalence of cannabis use has narrowed over time, partly due to cannabis liberalization.48 Changing social norms,8,9 availability of more potent cannabis products in the market,5-7 and cannabis legalization2 may contribute to shifting patterns of CUD and cannabis poisoning rates among males and females.
Moreover, our results show that adjusted prevalence rates of CUD and cannabis poisoning increased throughout the study period among those with MC and FFS insurance plans. Those with MC experienced a higher increase in the adjusted prevalence rates of CUD and cannabis poisoning than those with FFS plans during the study period, resulting in a 1.70-fold and a 6.62-fold increase, respectively. Although, to our knowledge, no previous data exist on cannabis use patterns among MC or FFS Medicaid enrollees, a recent study using Medicare data showed a rapid increase in the annual rate of health care encounters involving cannabis-related disorders among enrollees with MC compared with those with FFS.49
Although examining the reasons for increasing rates of CUD and cannabis poisoning in this Medicaid study sample is outside the scope of this study, the literature provides a few prospects that warrant further investigation. Possible explanations provided in the literature are cannabis liberalization,10,12,14 increased social acceptance of cannabis as harmless,8,9,32 availability of high-potency cannabis products in the market,5-7 and increased use of cannabis for self-treatment for various health conditions, such as anxiety and pain.50
Limitations and strengths
This study has a few limitations. First, this study uses insurance claims data from Medicaid enrollees living in multiple states and does not represent all Medicaid enrollees in the United States. Second, as our sample does not have state identifier information, we were unable to account for differences in CUD and cannabis poisoning rates in states with and without legalized cannabis. Third, it is possible that CUD and cannabis poisoning incidences were underdiagnosed in claims data. However, given the nature of claims data we are unable to detect if underdiagnosis is taking place or if underdiagnosis is driven by patient demographic or clinical characteristics. Following the literature, we used CUD and cannabis poisoning diagnoses identified by ICD-9-CM and ICD-10-CM codes as proxies to estimate the prevalence of CUD and cannabis poisoning. Fourth, multiple factors, such as the transition from ICD-9-CM to ICD-10-CM, could have affected the trends of CUD and cannabis poisoning, although this is not visible in the age, sex, and insurance-type adjusted trends data we present.
Despite these limitations, this study has several strengths. To our knowledge, this is the first study to examine trends in the prevalence of CUD and cannabis poisoning in Medicaid enrollees. This study uses a very large Medicaid claims database spanning 12 years. To our knowledge, this is also the first study to examine trends in CUD and cannabis poisoning by demographic characteristics in Medicaid—sex, age, race/ethnicity, and insurance type.
Conclusion
The adjusted prevalence rates of CUD and cannabis poisoning among Medicaid enrollees increased significantly during the 2011–2022 period. The findings underscore the significant increases in cannabis poisoning rates among 0–17-year and 65+-year age groups, females, and MC enrollees. Increasing rates of CUD and cannabis poisoning may become a burden for health care systems. Therefore, the public and health care providers should be made aware of the risks associated with cannabis use, while strategies to reduce increasing rates of CUD and cannabis poisoning must be introduced and implemented by policymakers. While policymakers may find it challenging with the changing social perceptions of cannabis,8,9 increased availability of high-potency products in the market,5-7 and state cannabis legalization,2 it is imperative to adopt a multipronged approach to reduce the increasing rates of CUD and cannabis poisoning, such as targeted education campaigns on safe use and storage of cannabis, school-based prevention programs, limits on THC (delta-9-tetrahydrocannabinol) potency, improving product labeling, and expanding Medicaid coverage for mental health and SUD treatment.
Supplementary Material
Supplementary material
Supplementary material is available at Health Affairs Scholar online.
Funding
This research was supported by the National Institutes of Health National Center for Advancing Translational Sciences through grant number UL1TR001998. Any opinion, findings, and conclusions or recommendations expressed in this manuscript are those of the authors and do not necessarily reflect the views of the National Institutes of Health.
Data availability
The data utilized in this article were provided by Merative by permission. The data cannot be shared publicly per the Data Use Agreement. Anyone interested in the data may request access from Merative via https://www.merative.com/documents/merative-marketscan-research-databases.