Marijuana use by middle-aged and older adults in the United States, 2015-2016

Abstract

Background:

Marijuana use is increasing among middle-aged and older adults in the US, but little is understood of its pattern of use by this population.

Methods:

We performed a cross-sectional analysis of responses from 17,608 adults aged ≥50 years from the 2015 and 2016 administrations of the National Survey on Drug Use and Health. Prevalence of past-year marijuana use was estimated and compared between middle-aged adults (age 50-64) and older adults (≥65). Characteristics of past-year marijuana users including demographics, substance use, chronic disease, and emergency room use, were compared to non-marijuana users and stratified by age group. Marijuana use characteristics were also compared between middle-aged and older adults. We used multivariable logistic regression to determine correlates of past-year marijuana use.

Results:

Prevalence of past-year marijuana use was 9.0% among adults aged 50-64 and 2.9% among adults aged ≥65. Prevalence of past-year alcohol use disorder (AUD), nicotine dependence, cocaine use, and misuse of prescription medications (i.e., opioids, sedatives, tranquilizers) were higher among marijuana users compared to non-users. In adjusted models, initiation of marijuana use <19 years of age [adjusted odds ratio (AOR)=13.43, 95% confidence interval (CI) 9.60, 18.78)], AUD (AOR=2.11, 95% CI 1.51, 2.94), prescription opioid misuse (AOR 2.49, 95% CI 1.61, 3.85), nicotine dependence (AOR=1.90, 95% CI 1.59, 2.26), and cocaine use (AOR 7.43, 95% CI 4.23, 13.03), were all associated with increased odds of past-year marijuana use.

Conclusion:

Marijuana use is becoming more prevalent in this population and users are also at high risk for other drug use.

1. Introduction

Attitudes towards marijuana use are changing considerably in the United States (U.S.) with a growing number of states legalizing medical and/or recreational marijuana (Gallup, 2018). Although current users are more likely to be young adults, the Baby Boomer generation is unique as it has had more experience with marijuana compared to any generation preceding them (Han and Moore, 2018). Prevalence of current marijuana use decreases with age (Substance Abuse and Mental Health Services Administration [SAMHSA], 2016a), but the Baby Boomer generation—which is now comprised of middle-aged and older adults often with prior experience with marijuana—is now experiencing large increases in recent use. A previous study using nationally representative data, from 2006 to 2013, found a 57.8% relative increase in past-year marijuana use among adults aged 50-64 and a 250% relative increase for adults aged 65 and older (Han et al., 2017).

Despite increases in recent marijuana use by older adults, few studies have examined correlates of use in this population, especially among adults with multiple chronic conditions. Marijuana may have benefits for older adults for neuropathic pain, spasticity, anorexia, and nausea and vomiting (Briscoe and Casarett, 2018). In addition, medical marijuana laws have shown to be associated with significant reductions in opioid prescribing for the Medicare Part D population (Bradford et al., 2018), which could reduce the use and risks of opioids for older adults. However, there may be acute and chronic health risks associated with marijuana use by older adults as well as interactions with prescribed medications (Han and Moore, 2018).

With more US adults supporting the legalization of marijuana (Gallup, 2018), it is imperative to understand the changing patterns of marijuana use by both middle-aged older adults who are more likely to have underlying chronic disease compared to younger adults. While data on the risks and benefits of marijuana use for middle-aged and older adults remain limited, identifying subgroups of older adults who may be at heightened risk for the adverse effects associated with marijuana use (e.g., through concomitant use of other drugs) is also important. Therefore, we used national data from the National Survey on Drug Use and Health (NSDUH) to provide updated estimates of the prevalence of marijuana use and to examine demographic and other drug use characteristics of its users among middle-aged and older adults.

2. Methods

2.1. Data Source and Study Population

Data from adults aged ≥50 surveyed in the 2015 and 2016 (n=17,608) NSDUH were analyzed. Analyses were limited to these two most recent cohorts as NSDUH revised much of its questionnaire in 2015, limiting our ability to examine trends over time (SAMHSA, 2018a). NSDUH is a cross-sectional survey of non-institutionalized individuals in the 50 US states and the District of Columbia. NSDUH obtained a nationally representative probability sample of individuals through four stages. Surveys were administered via computer-assisted interviewing (conducted by an interviewer) and audio computer-assisted self-interviewing (ACASI). Sample weights were provided by NSDUH to address unit- and individual-level non-response. Additional information on sampling and survey methods can be found elsewhere (SAMHSA, 2018b) The weighted interview response rates for 2015 and 2016 were 69.7% and 68.4%, respectively.

2.2. Measures

Participants were asked if they had ever used marijuana in their lifetime, and those reporting lifetime use were also asked about recency of use (i.e., past-year, past-month) and age of first use, which we coded into quartiles (i.e., age <16, 17-18, 19-21, >21) for descriptive purposes. Given the literature on prior use influencing future use of marijuana among older adults (Han and Moore, 2018), we also coded a separate variable indicating whether marijuana was initiated at an early age. Specifically, we created a trichotomous variable indicating whether participants reported ever initiating marijuana at age 18 or younger or at 19-21 years of age. These cutoffs were the 50^th and 75^th percentile cutoffs, respectively. The comparison group consisted of never-users and those whose first marijuana use was after the age of 21. Past-year users were asked if a doctor has recommended marijuana in the past year, and they were also asked about frequency of past-year use. Participants were also asked how much people risk harming themselves physically and in other ways when they smoke marijuana 1) once a month, and 2) once or twice a week. Answer options were no risk, slight risk, moderate risk, and great risk; however, we combined moderate and great risk due to the low prevalence of self-reported great risk.

Regarding other drug use, NSDUH asked about past-year use of a variety of drugs; we focused on past-year use of cocaine, and past-year misuse of prescription opioids, tranquilizers (benzodiazepines and muscle relaxants), and sedatives (which includes zolpidem, eszopiclone, and zaleplon products; the benzodiazepines: flurazepam, temazepam, and triazolam; and barbiturates). As of 2015, NSDUH defines misuse as using a drug in any way not directed by a doctor, including use without a prescription, more often, in greater amounts, or longer than the participant was directed to take them, or use in any other way a doctor did not direct the participant to use them. Nicotine dependence was assessed and defined based on dependence criteria of the Nicotine Dependence Syndrome Scale (NDSS) (Shiffman et al., 2004) and alcohol use disorder (AUD) was determined by responses to a series of questions determining if criteria were met for abuse or dependence as per Diagnostic and Statistical Manual of Mental Disorders, 4^th Edition (DSM-IV) criteria (American Psychiatric Association, 1994).

NSDUH categorized older adults into age 50-64 and age ≥65. Participants were also asked their gender, race/ethnicity, annual family income, and marital status. They were also asked questions to determine if they met DSM-IV criteria for a major depressive episode in the past year and whether they have been admitted to an emergency department (ED) for any reason in the past year. Finally, they were asked if they had ever been informed by a doctor or other medical professional that they have ever had the following 10 medical diseases: heart disease, diabetes, chronic obstructive pulmonary disease (COPD), cirrhosis of the liver, hepatitis B or C, kidney disease, asthma, HIV/AIDS, hypertension, and cancer. To examine medical multimorbidity, we further coded these indicators into ≥2 chronic conditions and ≥3 chronic conditions as has been performed in other studies (Swartz and Jantz, 2014).

2.3. Statistical Analysis

We first compared demographic, health, and drug use characteristics between participants age 50-64 and age ≥65. We then compared each characteristic according to whether past-year marijuana use was reported, stratified by age, and then within past-year marijuana users we compared marijuana-specific characteristics by age. All comparisons were conducted using chi-square. Finally, we computed bivariable and multivariable logistic regression models to examine how each characteristic is associated with past-year marijuana use as an outcome variable. Specifically, we computed unadjusted odds ratios (ORs) for each separate covariate, and then we fit all covariates (survey year, demographic characteristics, substance use, multimorbidity, depression, and all-cause ED use) simultaneously into a multivariable model which produced adjusted ORs (aORs) for all covariates.

For bivariable and multivariable analyses, we aggregated data for all adults over the age of 50 and adjusted for age. Analyses were conducted using Stata SE 13 (StataCorp, College Station, TX, 2013) weighted to account for the complex survey design, and used imputation-revised variables to limit missing data. Taylor series estimation methods were utilized to provide accurate standard errors (Heeringa et al., 2010). This secondary data analysis was exempt for review by New York University’s Institutional Review Board.

3. Results

Weighed sample characteristics are shown in Table 1. Prevalence of past-year marijuana use was 9.0% for adults aged 50-64 and 2.9% for adults aged ≥65, and prevalence of past-month use was estimated to be 5.7% for adults aged 50-64 and 1.7% for adults aged ≥65. More than half (54.5%) of adults aged 50-64 are estimated to have ever used marijuana with over a fifth (22.4%) of adults aged ≥65 estimated to have ever used.

Table 1.

Marijuana use and sample characteristics, 2015-2016^a

Characteristic	Age 50-64 (n=10,398b)	Age 65+ (n=7,210b)	p-value
Marijuana use
Marijuana ever used	54.6 (53.3, 55.8)	22.4 (21.0, 24.0)	<0.001
Marijuana past-year use	9.0 (8.2, 9.9)	2.9 (2.3, 3.5)	<0.001
Marijuana past-month use	5.7 (5.1, 6.4)	1.7 (1.4, 2.2)	<0.001
Marijuana dependence or abuse past-year	0.5 (0.4, 0.7)	0.1 (0.0, 0.2)	<0.001
Survey Year
2015	50.0 (48.6, 51.5)	49.3 (47.4, 51.2)	0.49
2016	50.0 (48.5, 51.4)	50.7 (48.8, 52.6)
Sex
Male	48.3 (47.0, 49.7)	44.8 (43.3, 46.3)	<0.001
Female	51.7 (50.3, 53.0)	55.2 (53.7, 56.7)
Race/ethnicity
Non-Hispanic White	69.9 (68.5, 71.3)	77.5 (76.0, 79.0)	<0.001
Non-Hispanic African American	11.4 (10.7, 12.2)	8.9 (8.2, 9.6)	<0.001
Hispanic	11.8 (11.0, 12.7)	8.1 (7.2, 9.2)	<0.001
Non-Hispanic Asian	4.7 (3.9, 5.6)	3.2 (2.6, 3.9)	<0.001
Other	2.2 (1.9, 2.5)	2.3 (2.0, 2.8)	<0.001
Total family income
<$20,000	15.2 (14.3, 16.2)	17.5 (16.3, 18.7)	<0.001
$20-$49,999	24.8 (23.7, 25.9)	37.5 (36.0, 39.1)	<0.001
$50,000-$74,999	16.8 (15.9, 17.7)	17.4 (16.1, 18.7)	<0.001
≥ $75,000	43.3 (41.9, 44.6)	27.6 (26.3, 29.0)	<0.001
Marital status
Married	63.5 (62.2, 64.7)	59.8 (58.1, 61.5)	<0.001
Widowed	5.1 (4.6, 5.6)	21.8 (20.6, 23.1)	<0.001
Divorced or separated	21.4 (20.4, 22.5)	14.0 (12.9, 15.3)	<0.001
Never married	10.0 (9.3, 10.7)	4.4 (3.9, 4.9)	<0.001
Unhealthy Alcohol use
Past-year alcohol use disorderc	4.9 (4.5, 5.4)	1.5 (1.2, 1.8)	<0.001
Tobacco Dependence
Nicotine dependenced	9.2 (8.5, 9.9)	3.5 (3.1, 4.0)	<0.001
Drug use
Past-year cocaine use	1.1 (0.8, 1.4)	0.1 (0.1, 0.2)	<0.001
Past-year prescription opioid misuse	3.5 (3.2, 4.0)	1.2 (0.9, 1.5)	<0.001
Past-year sedative misuse	0.5 (0.4, 0.7)	0.4 (0.3, 0.6)	0.41
Past-year tranquilizer misuse	1.5 (1.3, 1.9)	0.6 (0.4, 0.9)	<0.001
Chronic Disease
Heart condition	11.5 (10.7, 12.4)	28.3 (27.0, 29.6)	<0.001
Diabetes	14.6 (13.8, 15.5)	22.1 (20.8, 23.4)	<0.001
Chronic obstructive pulmonary disease	5.8 (5.3, 6.5)	9.4 (8.5, 10.3)	<0.001
Cirrhosis	0.6 (0.4, 0.8)	0.6 (0.5, 0.9)	0.59
Hepatitis B or C	2.5 (2.2, 3.0)	2.0 (1.6, 2.5)	0.15
Kidney disease	2.2 (1.9, 2.6)	5.1 (4.5, 5.8)	<0.001
Asthma	7.8 (7.2, 8.4)	7.0 (6.4, 7.7)	0.11
HIV/AIDS	0.3 (0.2, 0.4)	0.1 (0.0, 0.2)	0.02
Hypertension	28.2 (27.3, 29.2)	41.4 (40.1, 42.8)	<0.001
Cancer	7.2 (6.6, 7.9)	17.5 (16.5, 18.5)	<0.001
2 or more of the above chronic disease	18.8 (17.8, 19.9)	35.5 (34.1, 36.9)	<0.001
3 or more of the above chronic disease	6.7 (6.2, 7.2)	16.0 (15.0, 17.1)	<0.001
Mental Health
Past year major depressive episodec	6.2 (5.6, 6.8)	2.9 (2.4, 3.5)	<0.001
Health care utilization
All-cause emergency department use past year	23.7 (22.7, 24.6)	28.9 (27.5, 30.4)	<0.001

^aData from the 2015 and 2016 US National Survey on Drug Use and Health (NSDUH)

^bAll percentages are weighted and percentages have been rounded and may not sum to 100; CI=Confidence interval

^cBased on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition

^dBased on the Nicotine dependence syndrome scale (NDSS)

Table 2 presents comparisons between marijuana users and non-users stratified by age group, and between age groups. For adults aged 50-64, past-year marijuana users were more likely to be male (60.0% vs. 47.2%, p<0.001), non-Hispanic white (73.4% vs. 69.6%, p=0.003), and have a family income <$20,000 (21.1% vs. 14.6%, p<0.001). They were also less likely to be married (47.5% vs. 65.1, p<0.001), more likely to use the ED (27.5% vs. 23.3%, p=0.01), and more likely to have AUD, nicotine dependence, use cocaine, and misuse prescription opioids, sedatives, and tranquilizers compared to non-users. Regarding chronic diseases, past-year marijuana users aged 50-64 were less likely to have diabetes, but more likely to have COPD, cirrhosis, hepatitis B or C, HIV/AIDS, and depression (ps<05). Among adults age ≥65, past-year marijuana users were more likely to be male (68.8% vs. 44.1%, p<0.001), have a family income <$20,000 (25.3% vs. 17.3%, p<0.04); less likely to be married (46.4% vs. 60.2, p<0.001), and more likely to have AUD, nicotine dependence, use cocaine, and misuse prescription opioids (ps<.05) compared to non-users. In terms of chronic diseases, past-year marijuana users aged ≥65 were less likely to have diabetes, but more likely to have hepatitis B or C and depression (ps<.05). Between marijuana users age 50-64 and ≥65, marijuana users age 65 years of age and older were more likely to be male (68.8% vs. 60.0%, p=0.03), be widowed (13.5% vs. 5.4%, p=0.01), and less likely to have nicotine dependence, use cocaine, and misuse prescription opioids compared to marijuana users aged 50-64 (ps<.05). Regarding chronic diseases, marijuana users ≥65 were more likely to have heart conditions, COPD, kidney disease, hypertension, cancer, and have multimorbidity compared to marijuana users aged 50-64 (ps<.05).

Table 2:

Marijuana use by age group, demographics, substance use, and health status by age group, 2015-2016^a

	Age 50-64			Age 65+
Characteristic	No Past-Year Marijuana Use % (95% CI)b (n=9,443)	Past-Year Marijuana Use % (95% CI)b (n=955)	p- value	No Past-Year Marijuana Use % (95% CI)b (n=6,996)	Past-Year Marijuana Use % (95% CI)b (n=214)	p- value	p-value (Age 50- 64 vs. Age 65+)
Sex
Male	47.2 (45.7, 48.6)	60.0 (56.2, 63.6)	<0.001	44.1 (42.6, 45.7)	68.8 (60.3, 76.2)	<0.001	0.03
Female	52.8 (51.4, 54.3)	40.0 (36.4, 43.8)		55.9 (54.3,57.4)	31.2 (23.8, 39.7)
Race/ethnicity
Non-Hispanic White	69.6 (68.1,71.0)	73.4 (68.7, 77.6)	0.003	77.4 (75.8, 79.0)	80.2 (72.9, 85.9)	0.11	0.33
Non-Hispanic African American	11.1 (10.4, 11.9)	14.5 (11.9, 17.4)		8.7 (8.1, 9.5)	13.1 (8.7, 19.2)
Hispanic	12.3 (11.5, 13.1)	6.9 (4.5, 10.6)		8.3 (7.3, 9.3)	3.9 (0.9, 15.1)
Non-Hispanic Asian	4.9 (4.2, 5.8)	2.5 (1.0, 5.9)		3.2 (2.7, 4.0)	0.2 (0.0, 1.1)
Other	2.1 (1.8, 2.4)	2.8 (1.9, 4.1)		2.3 (2.0, 2.8)	2.7 (1.1, 6.4)
Total family income
<$20,000	14.6 (13.7, 15.6)	21.1 (18.1,24.3)	<0.001	17.3 (16.1, 18.5)	25.3 (18.7, 33.3)	0.04	0.02
$20-$49,999	24.6 (23.6, 25.7)	26.2 (22.6, 30.1)		37.7 (36.2, 39.2)	31.6 (23.8, 40.6)
$50,000-$74,999	16.9 (16.0, 17.8)	15.3 (12.4, 18.6)		17.3 (16.1, 18.6)	21.1 (14.0, 30.6)
≥ $75,000	43.9 (42.5, 45.2)	37.5 (33.4, 41.8)		27.8 (26.4, 29.2)	22.0 (16.4, 28.8)
Marital status
Married	65.1 (63.8, 66.3)	47.5 (43.8, 51.1)	<0.001	60.2 (58.5, 61.8)	46.4 (37.4, 55.7)	<0.001	0.01
Widowed	5.1 (4.6, 5.6)	5.4 (3.9, 7.5)		22.1 (20.9, 23.3)	13.5 (8.9, 19.8)
Divorced or separated	20.4 (19.3, 21.6)	31.7 (28.7, 34.8)		13.5 (12.4, 14.7)	30.8 (23.1, 39.8)
Never married	9.5 (8.8, 10.2)	15.4 (13.0, 18.2)		4.2 (3.7, 4.8)	9.3 (5.5, 15.2)
Unhealthy Alcohol use
Past-year alcohol use disorderc	4.0 (3.6, 4.4)	14.2 (11.2, 17.9)	<0.001	1.3 (1.0, 1.6)	8.2 (4.4, 14.7)	<0.001	0.08
Tobacco Dependence
Nicotine dependenced	7.7 (7.1, 8.3)	23.9 (20.6, 27.5)	<0.001	3.3 (2.9, 3.8)	10.7 (6.7, 16.8)	<0.001	<0.001
Early Marijuana Use
Never Used or First Used After Age 21	55.9 (54.6, 57.2)	7.1 (5.1, 9.8)	<0.001	90.9 (90.1, 91.7)	45.3 (36.1, 54.8)		<0.001
First Used before age 19	35.4 (34.1, 36.7)	82.7 (79.3, 85.6)		3.6 (3.1, 4.1)	24.7 (17.7, 33.4)
First Used age 19-21	8.7 (8.1, 9.5)	10.3 (8.2, 12.8)		5.5 (5.0, 6.2)	30.0 (23.1, 37.9)
Drug use
Past-year cocaine use	0.4 (0.3, 0.6)	8.0 (6.1, 10.4)	<0.001	0.3 (0.0, 0.1)	2.9 (1.1, 7.2)	<0.001	0.04
Past-year prescription opioid misuse	2.7 (2.4, 3.2)	11.6 (9.3, 14.3)	<0.001	1.1 (0.8, 1.4)	5.1 (2.3, 10.9)	<0.001	0.04
Past-year sedative misuse	0.4 (0.3, 0.6)	1.2 (0.6, 2.4)	0.02	0.4 (0.2, 0.6)	1.5 (0.5, 4.5)	0.02	0.69
Past-year tranquilizer misuse	1.2 (1.0, 1.5)	4.8 (3.3, 7.0)	<0.001	0.5 (0.4, 0.8)	3.2 (1.3, 7.6)	<0.001	0.41
Chronic Disease
Heart condition	11.4 (10.6, 12.3)	12.5 (10.0, 15.5)	0.45	28.4 (27.0, 29.7)	25.3 (17.9, 34.5)	0.49	<0.001
Diabetes	14.9 (14.1, 15.8)	11.6 (9.3, 14.3)	0.02	22.4 (21.1, 23.7)	12.6 (8.5, 18.4)	0.003	0.68
Chronic obstructive pulmonary disease	5.7 (5.1, 6.3)	7.6 (5.9, 9.9)	0.04	9.3 (8.4, 10.2)	12.6 (8.4, 18.7)	0.15	0.05
Cirrhosis	0.5 (0.4, 0.7)	1.2 (0.6, 2.5)	0.02	0.6 (0.5, 0.9)	0.1 (0.0, 0.9)	0.07	0.01
Hepatitis B or C	2.2 (1.8, 2.5)	6.4 (4.7, 8.6)	<0.001	1.9 (1.6, 2.4)	6.1 (2.6, 13.4)	0.005	0.92
Kidney disease	2.3 (1.9, 2.7)	1.8 (1.1, 3.0)	0.37	5.1 (4.4, 5.8)	5.7 (2.8, 11.6)	0.74	0.01
Asthma	7.9 (7.2, 8.5)	6.8 (4.7, 9.7)	0.44	7.0 (6.4, 7.7)	7.8 (3.9, 15.2)	0.74	0.73
HIV/AIDS	0.2 (0.1, 0.3)	1.0 (0.5, 2.1)	<0.001	0.1 (0.0, 0.2)	0.0 (0.0, 0.0)	0.73	0.20
Hypertension	28.2 (27.2, 29.2)	28.3 (25.2, 31.6)	0.96	41.6 (40.1, 43.0)	37.0 (29.7, 44.9)	0.27	0.02
Cancer	7.3 (6.7, 7.9)	7.0 (5.3, 9.2)	0.77	17.4 (16.4, 18.5)	19.8 (13.0, 29.1)	0.54	<0.001
2 or more of the above chronic disease	18.6 (17.5, 19.7)	21.2 (18.2, 24.6)	0.11	35.5 (34.0, 37.0)	33.8 (26.8, 41.5)	0.66	<0.001
3 or more of the above chronic disease	6.6 (6.1, 7.1)	7.7 (5.9, 9.9)	0.24	15.9 (14.9, 17.0)	17.7 (12.1, 25.0)	0.58	<0.001
Mental Health
Past year major depressive episodec	5.8 (5.2, 6.5)	10.0 (8.1, 12.3)	<0.001	2.8 (2.3, 3.4)	7.3 (4.0, 12.8)	0.003	0.28
Health care utilization
All-cause emergency department use past year	23.3 (22.3, 24.3)	27.5 (24.4, 30.8)	0.01	29.0 (27.6, 30.4)	27.0 (18.4, 37.8)	0.69	0.93

^aData from the 2015 and 2016 US National Survey on Drug Use and Health (NSDUH)

^bAll percentages are weighted and percentages have been rounded and may not sum to 100; CI=Confidence interval

^cBased on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition

^dBased on the Nicotine dependence syndrome scale (NDSS)

Table 3 presents the use characteristics and perceived risk of marijuana among adults who reported past-year marijuana use by age group. Among adults aged ≥65 with past-year marijuana use, 22.9% were recommended to take it by a doctor, compared to 15.0% for adults 50-64 although this difference was not significant (p=0.11). The majority (82.7%) of those aged 50-64 reported using marijuana for the first time at or before age 19, and a quarter (24.5%) of those age ≥65 reported using at or before age 19 (p<.001). The frequency of reported marijuana use had two peaks for both 50-64-year-olds and those aged ≥65: reported use 1 to 11 days and 100-299 days. Most marijuana users in both age groups reported no risk or slight risk for using marijuana monthly or once or twice a week.

Table 3.

Marijuana use characteristics by age group, 2015-2016^a

Characteristic	Age 50-64 with past-year marijuana useb (n=955)	Age 65 and older with past-year marijuana useb (n=214)	p-value
Medical marijuana
Any marijuana use recommended by doctor in past 12 months	15.0 (11.5, 19.4)	22.9 (14.1, 34.9)	0.11
Age of initiation of marijuana use, years
<16	60.1 (55.7, 64.3)	7.5 (4.5, 12.3)	<0.001
17-18	22.6 (19.1, 26.5)	17.2 (11.3, 25.3)
19-21	10.3 (8.2, 12.8)	30.0 (23.1, 37.9)
>21	7.1 (5.1, 9.8)	45.3 (36.1, 54.8)
Frequency of marijuana use in the past 12 months, in days
1 to 11	30.2 (27.1, 33.6)	35.2 (29.0, 42.1)	0.14
12 to 49	16.8 (14.6, 19.3)	17.2 (11.4, 25.1)
50 to 99	12.2 (9.6, 15.3)	6.4 (3.6, 11.3)
100 to 299	26.1 (22.8, 29.8)	31.5 (22.6, 42.1)
300+	14.7 (12.4, 17.3)	9.7 (5.6, 16.1)
Perceived risk of cannabis use once a month
No risk	51.6 (47.9, 55.2)	59.4 (49.8, 68.3)	0.19
Slight risk	34.5 (31.0, 38.1)	31.9 (24.9, 40.0)
Great/Moderate risk	14.0 (11.3, 17.1)	8.7 (4.7, 15.7)
Perceived risk of cannabis use once or twice a week
No risk	43.9 (40.8, 8.5)	52.7 (42.3, 62.8)	0.31
Slight risk	40.9 (37.0, 44.9)	34.6 (25.3, 45.3)
Great/Moderate risk	15.3 (12.5, 18.4)	12.7 (7.1, 21.8)

^aData from the 2015 and 2016 US National Survey on Drug Use and Health (NSDUH)

^bAll percentages are weighted and percentages have been rounded and may not sum to 100; CI=Confidence interval

As shown in Table 4, which presents correlates of past-year marijuana use, older participants were at low odds for reporting marijuana use (OR=0.30, p<.001), but this association was lost in the multivariable model. Females were consistently at lower odds for use than males (AOR=0.72, p=0.002), and although compared to white participants, black participants were at higher odds for use (OR=1.38, p=0.01) and Hispanic (OR=0.59, p=0.01) participants were at lower odds for use, these race associations disappeared in the multivariable model. Likewise, higher incomes were at lower odds for use (incomes ≥ $75,000: OR=0.68, p=0.001), but was not significant in the multivariable model. In the multivariable model, past-year AUD (AOR=2.11, p<001), nicotine dependence (AOR=1.90, p<001), cocaine use (AOR=7.43, p<001), and opioid misuse (AOR=2.49, p<001) were all consistent risk factors for use, as was reporting using marijuana before age 22 (for <19: AOR=13.43, p<.001; for 19-21: AOR=8.82, p<.001). In a sensitivity analysis, we stratified bivariable and multivariable models separately by age (50-64 and 65 and older) and found similar patterns. Past-year sedative and tranquilizer use as well as having a major depressive episode in the past year, increased odds for marijuana use until controlling for all other covariates.

Table 4.

Correlates of past-year marijuana by older adults, 2015-2016 (n=17,608)

Characteristic	Unadjusted Odds Ratio	(95% CI)	p- value	Adjusted Odds Ratioa	(95% CI)	p- value
Survey Year
2015	1.00			1.00
2016	1.01	(0.86, 1.19)	0.88	0.99	(0.82, 1.19)	0.94
Age
Age 50-64	1.00			1.00
65+	0.30	(0.24, 0.36)	<0.001	1.13	(0.88, 1.47)	0.33
Sex
Male	1.00			1.00
Female	0.53	(0.44, 0.63)	<0.001	0.72	(0.59, 0.89)	0.003
Race/ethnicity
Non-Hispanic White	1.00			1.00
Non-Hispanic African American	1.38	(1.10, 1.73)	0.01	1.16	(0.88, 1.52)	0.29
Hispanic	0.59	(0.39, 0.88)	0.01	0.83	(0.51, 1.35)	0.45
Non-Hispanic Asian	0.48	(0.20, 1.11)	0.08	1.34	(0.56, 3.22)	0.51
Other	1.23	(0.85, 1.80)	0.27	1.12	(0.75, 1.66)	0.58
Total family income
<$20,000	1.00			1.00
$20-$49,999	0.65	(0.52, 0.79)	<0.001	0.88	(0.68, 1.13)	0.31
$50,000-$74,999	0.69	(0.56, 0.87)	0.002	0.89	(0.69, 1.16)	0.39
≥ $75,000	0.68	(0.55, 0.84)	0.001	0.84	(0.63, 1.13)	0.26
Marital status
Married	1.00				1.00
Widowed	0.73	(0.54, 0.99)	0.05	1.31	(0.94, 1.84)	0.11
Divorced or separated	2.42	(2.11,2.78)	<0.001	1.65	(1.35,2.01)	<0.001
Never married	2.66	(2.15, 3.28)	<0.001	1.68	(1.25,2.25)	0.001
Unhealthy Alcohol use
Past-year alcohol use disorderb	5.28	(3.97, 7.02)	<0.001	2.18	(1.57, 3.03)	<0.001
Tobacco Dependence
Nicotine dependencec	4.46	(3.78, 5.27)	<0.001	1.87	(1.57, 2.23)	<0.001
Early marijuana use
Marijuana use <19 years of age	16.75	(12.83, 21.87)	<0.001	13.43	(9.60, 18.78)	<0.001
Marijuana use between ages 19-21	9.54	(7.23, 12.59)	<0.001	8.82	(6.43, 12.10)	<0.001
Drug use
Past-year cocaine use	31.54	(20.66, 48.16)	<0.001	7.73	(4.33, 13.79)	<0.001
Past-year prescription opioid misuse	5.65	(4.27, 7.47)	<0.001	2.49	(1.60, 3.88)	<0.001
Past-year sedative misuse	3.07	(1.52,6.21)	0.002	0.97	(0.41, 2.24)	0.95
Past-year tranquilizer misuse	5.08	(3.40, 7.60)	<0.001	1.14	(0.63, 2.06)	0.65
Multimorbidity
2 or more chronic diseasesd	0.87	(0.73, 1.05)	0.14	1.05	(0.84, 1.33)	0.65
Mental Health
Past year major depressive episode	0.44	(0.35, 0.57)	<0.001	0.81	(0.56, 1.16)	0.24
Health care utilization
All-cause emergency department use past year	1.08	(0.92, 1.28)	0.34	0.93	(0.77, 1.13)	0.45

^aMultivariable model adjusted for all variables

^bBased on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition¹⁰

^cBased on the Nicotine dependence syndrome scale (NDSS)⁹

^dChronic conditions include: Asthma, Chronic obstructive pulmonary disease, Cirrhosis, Diabetes, Heart Disease, Hepatitis, High Blood Pressure, HIV/AIDS, Cancer, and kidney disease

4. Discussion

Using the most recent data from a nationally representative survey of older adults in the US, we estimate that 9.0% of adults aged 50-64 and 2.9% of adults aged ≥65 used marijuana in the past year. Thus, use may be increasing among this older population as prevalence appears to be higher compared to an earlier study that found that 7.1% of adults aged 50-64 and 1.4% of adults aged ≥65 used marijuana in 2012-2013 (Han et al., 2017). Utilizing new survey questions added to NSDUH in 2013, we found that a quarter of marijuana users age ≥65 reported that a doctor had recommended marijuana in the past year.

A concerning finding from our study was the higher prevalence of AUD, nicotine dependence, cocaine use, and prescription drug misuse among middle-aged and older adults with past-year marijuana use compared to non-past-year users. While our study could not determine whether these drugs were used concomitantly, these results suggest that there may be a population of middle-aged and older adults who potentially engage in unsafe polysubstance use. Several studies have noted the additive adverse effects of the simultaneous use of marijuana and alcohol on cognitive and motor functioning (Chesher et al., 1976; Hartman et al., 2015; Kelly et al., 2004). Our findings highlight the importance to screen older patients who use marijuana for other substance use to ensure patients are educated about the potential risks of co-use. While other studies have highlighted the association between medical marijuana laws and decreases in both opioid prescribing⁶ and opioid overdoses (Bachhuber et al., 2014), we found a significant positive association between past-year marijuana use and past-year prescription opioid misuse. These results are similar to other studies that found an association between medical marijuana use and nonmedical use of prescription drugs including opioids and opioid use disorder (Olfson et al., 2018; Camputi and Humphreys, 2018).

Further studies, especially among older adults, are needed to understand better the associations between marijuana use and subsequent patterns of prescription opioid use as it is unknown whether marijuana is being used as an opioid replacement or whether one drug is merely a risk factor for the use of the other drug. In addition, one study has shown a large proportion of older medical marijuana users also use recreationally and have higher rates of marijuana use disorder compared to recreational users (Choi et al., 2017a). Therefore, older adults who use marijuana, including for medical purposes, should also specifically be screened for both unhealthy marijuana use and prescription opioid use and misuse.

Our study is the first to our knowledge to describe correlates of marijuana use by chronic disease. While it may be expected that adults with multimorbidity would be more inclined to use marijuana for symptom management compared to adults without chronic diseases (Briscoe and Casarette, 2018; Han and Moore, 2018), in our study, only hepatitis B or C were higher among marijuana users in both the age 50-64 and age ≥65 older groups compared to non-users. Multimorbidity was not significantly higher among marijuana users or correlated with marijuana use in our adjusted models. Demographic correlates of past-year marijuana use including males and not being married are consistent with findings in previous studies (Choi et al., 2016; Han et al., 2017). Also consistent with previous literature is the finding that older adults who use marijuana tended to have used when they were much younger, with nearly all adults age 50-64 and more than half of adults age ≥65 first using marijuana when they were 21 years of age or younger. Use before age 22 was also strongly correlated with past-year marijuana use. This likely reflects the boom in popularity of marijuana in the 1960s and 1970s and emphasizes that the Baby Boomer generation has had more experience with marijuana compared to earlier generations (Han and Moore, 2018). Most of these older users are not new initiates although more research is needed to examine whether there are demographic, drug use, and health-related differences according to whether such older individuals used continuously or sporadically through life or have recently reinitiated. However, it is important to note that the potency of marijuana (tetrahydrocannabinol [THC]) concentration may have increased considerably over the past several decades (Choi et al., 2016; ElSohly et al., 2000; Slade et al., 2012). Therefore, middle-aged and older adults who used marijuana before age 22 may be more familiar with lower potency marijuana than what is available today. This is particularly important as aging is associated with physiological changes and coupled with stronger potency marijuana, could pose a higher risk associated with marijuana use in this population.

This study also found that among marijuana users, many perceived no risk of harm associated with use. A previous study has found among older marijuana users a higher proportion that perceived no to slight risk of marijuana compared to never-users, although those with marijuana use disorder were more likely to report moderate or great risk perception of marijuana (Choi et al., 2017b). The latter finding, the authors postulated, was because adults with marijuana use disorder may have experienced the negative effects of marijuana and therefore better understand its risks while occasional users who do not perceive risk may not have experienced them. While marijuana use may be relatively safer than other psychoactive substances (Nutt et al., 2007), the risks of its use still need to be evaluated. It is also unknown how “risk” was interpreted as legal risk is more of an issue in many states than in others.

In comparison, alcohol use among older adults has been studied to assess better specific risks, potential benefits, interactions with prescribed medications, and the creation of guidelines to quantify safer alcohol levels for older adults (Han and Moore, 2018; National Institute on Aging, 2017). However, little research has examined marijuana use in the same way, but some studies have pointed out risks that are relevant for older adults. Acute adverse effects of marijuana use can include anxiety, dry mouth, tachycardia, high blood pressure, palpitations, wheezing, confusion, and dizziness (Seamon et al., 2007; Volkow et al., 2014). As such, older marijuana users have been found to have a higher rate of injury and ED visits compared to non-users (Choi et al., 2017c), although in our study we only found higher ED usage among marijuana users age 50 to 64 compared to non-users the same age range Chronic use can lead to chronic respiratory conditions, depression, impaired memory, and reduced bone density (Seamon et al., 2007; Volkow et al., 2014). Both cannabidiol (CBD) and THC compounds found in marijuana have effects on cytochrome P450 enzymes (Yamaori et al., 2011; Yamaori et al., 2012), thereby potentially affecting the metabolism of many classes of medications (Kelly et al., 2004; Yamreudeewong et al., 2009). Patients need to be informed of the possible adverse effects of marijuana, including potential drug-drug interactions, and that studies have not yet fully delineated the risks and benefits for older adults.

This study has several limitations. The NSDUH relies on self-report and is subject to recall and social desirability bias. The latter may be particularly true for substance use where respondents may deny their substance use, although the survey attempts to limit this via ACASI. Also, recall bias may be an issue for older adults to recall events, particularly from the distant past. The NSDUH also samples only the non-institutionalized US population. Finally, the survey is cross-sectional, and different participants were sampled each year; therefore, this study cannot establish causality. We were also unable to distinguish between marijuana users who may have been using continuously or intermittently for many years versus users who may have used decades ago and recently reinitiated use.

5. Conclusions

Our study characterizes the correlates of marijuana use by middle-aged and older adults and identifies a subgroup of older marijuana users who also engage in potentially risky unhealthy substance use. This emphasizes the importance of screening this population for polysubstance use. While the prevalence of marijuana use in this population will likely continue to increase, the risks and benefits are still unclear, and providers need to inform their older patients of the lack of current research on effects of use in this population.

Highlights.

• Estimates of past-year marijuana use has increased for middle-aged and older adults

• Older marijuana users were more likely to use other substances compared to non-users

• Screening for concurrent substance use among older marijuana users is an imperative