Abstract
Aims
Substance use is a significant public health problem among women in the United States, although limited data exist on patterns of use across different demographic groups. This cross-sectional survey study aimed to examine in adult women (1) patterns of substance use (i.e., alcohol, cannabis, other drugs, and prescription medication misuse) and (2) whether these patterns varied by age cohort.
Methods
From July to September 2024, we conducted an online cross-sectional survey of women ages 18–70 in the community who reported substance use in the prior 30 days. Using latent class analysis, we analyzed data from 2296 women.
Results
Overall, heavy use was common particularly for alcohol and cannabis (approximately 18% and 22%, respectively). Heavy substance use patterns varied qualitatively with age. The prevalence of heavy alcohol use (≥16 days/month) increased with older age cohort, as did prescription medication misuse. Heavy cannabis use (≥ 16 days/month) decreased with older age cohort. In a latent class analysis of substance use patterns, four discrete classes emerged: (1) alcohol use with some cannabis (approximately 57% of the sample), (2) heavy cannabis with some alcohol (27%), (3) polysubstance use (10%), and (4) prescription medication misuse with some alcohol (6%).
Conclusions
Among adult women who reported recent substance use, more than half used more than one substance, and use patterns varied with age cohort. Consideration of polysubstance use, age cohort, and life stage is critical in designing effective substance use prevention, early intervention, and treatment for adult women.
Keywords: Women, Female, Polysubstance, Substance use, Substance use disorders, Cannabis, Alcohol
Graphical abstract
Highlights
-
•
Survey of 2296 women's substance use examined differences across age cohorts 18–70.
-
•
Latent class analysis shows 4 distinct patterns of substance use that vary with age.
-
•
Alcohol use and prescription medication misuse (≥ 16 days/month) increased with age.
-
•
Cannabis use and other drug use (≥ 16 days/month) decreased with age.
-
•
Polysubstance use is common with >50% using more than one substance.
1. Introduction
In the past three decades, the gender gap in the prevalence of substance use disorders (SUDs) in the United States has dramatically narrowed (Grucza et al., 2008; Keyes et al., 2008; Keyes et al., 2010; McHugh et al., 2021). This includes alcohol, opioids, and cannabis, with rising rates in women and girls across domains of race, ethnicity, and lifespan (McHugh et al., 2021; Substance Abuse and Mental Health Services Administration, 2021). These trends are concerning, as women tend to have more severe substance-related medical and social consequences and most women do not seek treatment for substance use problems (Greenfield et al., 2010; Kay et al., 2010; National Institute on Alcohol Abuse and Alcoholism, 2025).
There is a critical need to understand patterns of substance use in women to improve understanding of risk patterns across the lifespan. While studies often track use patterns for specific substances (e.g., alcohol, cannabis, opioids), the use of multiple substances (i.e., polysubstance use) is the norm among individuals who use substances (Rockhill et al., 2025), including among those with mild substance use severity (Bailey & McHugh, 2023). Currently, little is known about use patterns among community samples of women in the U.S., especially how these patterns may vary across the lifespan (Bailey & McHugh, 2023; Rockhill et al., 2025).
As part of a larger project to identify correlates of and potential risk factors for heavy substance use in women, this study aimed to elucidate substance use patterns in adult women who reported using substances in the past 30 days to answer the following questions:
-
1.
What are the patterns of use of substances, including alcohol, cannabis, other drugs, and prescription medication misuse, in a community sample of adult women who have used substances in the past 30 days?
-
2.
Do these patterns of use vary across four age cohorts (18–30 years, 31–43, 44–56, and 57–70)?
2. Methods
2.1. Participants
Participants were community members recruited from online panels managed by Sago Health (Iselin, NJ) (Sago, 2024). Invitations were sent by Sago to individuals within the target age range. Those interested in participating were directed to click a link to a screening form to determine eligibility. Inclusion criteria were: 1) self-identified as female on Sago panel profile, 2) age 18–70 years, and 3) any reported substance use during the prior 30 days. To ensure adequate representation of both low- and high-frequency substance use, after the first 1495 respondents, the substance use inclusion criterion was changed to endorsement of at least 6 days in the past month in at least one of the four substance categories. After an additional 3073 respondents were screened, this inclusion criterion was reverted to the original for the remaining 337 respondents. There were no exclusion criteria.
Of the 4876 participants assessed for eligibility, 2405 (49.3%) were determined ineligible due to no substance use (n = 1385) and mild substance use only (n = 1020) in the second phase of survey recruitment (see Fig. 1). One hundred and seventy-four eligible participants were directed to the Information Sheet but chose not to participate. Of the remaining 2297 participants, one was identified as spam and removed from all analyses. A total of 2296 individuals participated in the survey.
Fig. 1.
Enrollment flow-chart.
2.2. Procedure
The study was determined to be exempt by the Mass General Brigham Institutional Review Board. After clicking the invitation link, potential participants completed four screening questions to assess their use of different substances (i.e., alcohol, cannabis, other drugs, and prescription medication misuse) over the prior 30 days. Eligible participants were directed to complete informed consent, and those who agreed to participate were directed to the survey. The survey was hosted on Research Electronic Data Capture (REDCap) (Harris et al., 2009; Harris et al., 2019). Data were collected between July and September 2024. Participants who completed the study were compensated 200 points (equivalent to $2) in their Sago account.
2.3. Measures
The survey included validated questionnaires to measure substance use and social and psychological characteristics. The following measures were used in these analyses.
Screener Questionnaire. The screener questionnaire asked about the frequency of substance use in the past 30 days, defined as any use of alcohol, cannabis, other drugs, and prescription medication misuse. Each screening question included four answer choices representing no use (“none”), mild use (“1–5 days”), moderate use (“6–15 days”), and heavy use (“16 or more days”).
Demographics and Social Determinants of Health. This instrument (designed for the study) assessed participant age, race, ethnicity, marital status, level of education, employment status, housing and financial statuses, household composition, substance use problems among family members, and other characteristics.
Tobacco, Alcohol, Prescription Medications, and Other Substances (TAPS) (McNeely et al., 2016). The TAPS-1 includes 4 items assessing use/misuse of tobacco, alcohol, prescription drugs, and other drugs (including marijuana) in the prior 12 months. Each substance or substance group is rated on a 5-point Likert scale ranging from “never” to “daily or almost daily.” TAPS-2 assesses substance use and related behaviors in the prior 3 months for tobacco cigarettes, alcohol, marijuana, cocaine, crack, and methamphetamine, heroin and fentanyl, misuse of prescription opiate pain relievers, misuse of medication for anxiety or sleep, misuse of medication for ADHD, and use of other illicit or recreational drugs.
2.4. Data analysis
First, we used descriptive statistics to characterize substance use behaviors overall and by age cohort. Next, latent class analysis (LCA) was conducted using the four response categories from the Screener Questionnaire assessing past 30-day frequency of alcohol, cannabis, other illicit substance use, and prescription medication misuse. LCA models were estimated in Mplus (Muthèn & Muthèn, 2017[1998]) utilizing a robust maximum likelihood estimation (“MLR” estimator in Mplus) that is appropriate for categorical indicators. A series of LCA models were fit (i.e., 2-class, 3-class, four-class…) and evaulated based on the following criteria: (1) theoretical interpretability, (2) Bayesian information criteria (BIC) with lower values indicating better model fit, (3) entropy, which is a metric that ranges from 0 to 1, with higher values indicating higher demarcation between classes, and (4) the size of identifed classes. There were no missing data in variables used in LCA. Participants were assigned to latent classes based on their highest estimated class membership probability. Figures were generated using ggplot (Wickham, 2016).
Multionmial regressions were estimated, in R statistical software version 4.4.3 (R Core Team, 2024) utilizing the VGAM package (Yee, 2010), to test the associations between age cohort and latent class membership. For the presented model, the youngest age cohort was used as the reference category. Models were also rerun with other age cohorts as the reference category to obtain all direct statistical comparisons.
3. Results
3.1. Demographics
Demographic data are summarized in Table 1. Briefly, participants had a mean age of 47.1 (SD = 13.3) years with the following age cohorts: 18–30 years (n = 319); 31–43 years (n = 642); 44–56 years (n = 655); and 57–70 years (n = 680). We divided the study age range into four equal-length age cohorts that correspond to four adult life stages of emerging adulthood (18–30 years), young-to-middle age adulthood (31–43), and middle- (44–56) to later- (57–70) periods of late adulthood (Halloran, 2024).
Table 1.
Demographics and social determinants of health for whole sample (n = 2296).
| Age in years [M (SD)] | 47.1 (13.3) |
|---|---|
| Age Cohort [n (%)] | |
| 18–30 years | 319 (13.9) |
| 31–43 years | 642 (28.0) |
| 44–56 years | 655 (28.5) |
| 57–70 years | 680 (29.6) |
| Ethnicity [n (%)] | |
| Hispanic/Latino | 265 (11.5) |
| Non-Hispanic/Non-Latino | 2023 (88.1) |
| Don't know/prefer not to answer | 8 (0.2) |
| Race [n (%)] | |
| White | 1659 (72.3) |
| Black/African-American | 429 (18.7) |
| Asian | 46 (2.0) |
| American Indian/Alaska Native | 22 (1.0) |
| Native Hawaiian/Other Pacific Islander | 2 (0.1) |
| Multi-Racial | 75 (3.3) |
| Other | 24 (1.0) |
| Don't know/prefer not to answer | 28 (1.2) |
| Marital Status [n (%)] | |
| Married | 950 (41.4) |
| Never married | 599 (26.1) |
| Divorced | 317 (13.8) |
| Cohabitating | 262 (11.4) |
| Widowed | 96 (4.2) |
| Separated | 62 (2.7) |
| Don't know/prefer not to answer | 10 (0.4) |
| Has children [n (%)]a | |
| Yes | 1447 (63.1) |
| No | 841 (36.6) |
| Prefer not to answer | 7 (0.3) |
| Number of children – entire sample [M (SD)] N = 2288 |
1.4 (1.4) |
| Number of children – of those who have children [M (SD)]b | 2.3 (1.2) |
| Children living in the household [n (%)] | |
| No | 1319 (57.4) |
| Yes | 977 (42.6) |
| Number of children in household [M (SD)] – entire sample | 0.8 (1.1) |
| Number of children in household – of those with children in the household [M (SD)]; n = 977 | 1.9 (1.0) |
| Sexual Orientation [n (%)] | |
| Heterosexual | 1957 (85.2) |
| Bisexual | 199 (8.7) |
| Gay/Lesbian | 54 (2.4) |
| Pansexual | 34 (1.5) |
| Queer | 18 (0.8) |
| Asexual | 13 (0.6) |
| Other | 3 (0.1) |
| Prefer not to answer | 18 (0.8) |
| Employment [n (%)] | |
| Working full-time | 1123 (48.9) |
| Working part-time | 379 (16.5) |
| Retired | 261 (11.4) |
| Unemployed | 148 (6.4) |
| Work within the home (without pay) | 141 (6.1) |
| Student | 62 (2.7) |
| Temporarily laid off, sick/maternity leave | 19 (0.8) |
| Disabled | 140 (6.1) |
| Other | 11 (0.5) |
| Prefer not to answer | 12 (0.5) |
| Education Level [n (%)] | |
| Less than high school | 19 (0.8) |
| High school graduate or equivalent | 273 (11.9) |
| Some college | 627 (27.3) |
| Associate's degree | 272 (11.8) |
| Bachelor's degree | 726 (31.6) |
| Graduate degree | 375 (16.3) |
| Prefer not to answer | 4 (0.2) |
| Housing [n (%)]a | |
| Apartment/house | 2249 (98.0) |
| College dormitory | 9 (0.4) |
| Group home, halfway house, or temporary or transitional housing | 17 (0.7) |
| Homeless shelter | 5 (0.2) |
| Outdoors/street, abandoned or public building, automobile | 13 (0.6) |
| Hospital or other treatment facility | 2 (0.1) |
| Needed medical treatment in the past year and didn't get it because of cost [n (%)] (n = 2103) | |
| Total sample | 607 (28.9) |
| 18–30 cohort | 109 (37.1) |
| 31–43 cohort | 190 (33.0) |
| 44–56 cohort | 192 (31.9) |
| 57–70 cohort | 116 (18.3) |
| Had enough income for necessities in the past month [n (%)] (n = 2112) | |
| None/almost none of the time | 243 (11.5) |
| Some of the time | 560 (26.5) |
| All/almost all of the time | 1309 (62.0) |
Note: Sample size is 2296 unless otherwise noted.
In this sample, 72.3% identified as white and 18.7% as Black/African American and 41.4% percent were married. Among the 63.1% who had children, the mean number of children was 2.3 (SD = 1.2) and 42.6% had children living in the household. Among participants, 28.9% reported being unable to afford the medical care they needed in the past year, and 38.0% did not have financial resources to cover basic living expenses all or some of the time in the past month.
3.2. Substance use descriptives by age
Heavy Use Patterns of Specific Substances: All participants used at least one substance in the past 30 days, per study inclusion criteria. Over this period, 84.8% of participants had used alcohol at least one day, with 18.3% endorsing heavy use (i.e., ≥ 16 days). Forty-five percent used cannabis, with 22.6% endorsing heavy use; 19.1% endorsed prescription medication misuse, with 8.6% misusing on ≥16 days; and 11.2% endorsed using other substances, with 6.1% endorsing heavy use. These patterns of heavy use (≥16 days in past 30 days) varied by age cohort (see Fig. 2). Qualitatively, heavy alcohol use (≥ 16 days) was lowest in the youngest age cohort and increased with each succeeding age cohort. This contrasted with heavy cannabis use, which was highest in the youngest age cohort and decreased with age. Prescription medication misuse of 16 or more days per month was highest in the two oldest cohorts; heavy use of other drugs was highest in two middle-aged cohorts. Table 2 summarizes use patterns for alcohol, cannabis, and tobacco for the total sample and by age cohort. Notably, among those endorsing drinking in the past 3 months, drinking 3 or more days per week was more prevalent in older age cohorts with 43.9% of those 57–70 years reporting drinking at that level compared with 30.8% of those 18–30 years.
Fig. 2.
Days of Use (none, mild, moderate and heavy) by age cohort for Alcohol, Cannabis, Prescription Medications, and Other Drugs.
Table 2.
Substance use by age cohort for alcohol, Cannabis, and Tobacco/Nicotine.
| Measure | Total sample |
18–30 cohort |
31–43 cohort |
44–56 cohort |
57–70 cohort |
|||||
|---|---|---|---|---|---|---|---|---|---|---|
| n (%) |
sample size | n (%) |
sample size | n (%) |
sample size | n (%) |
sample size | n (%) |
sample size | |
| Alcohol | ||||||||||
| Any use in past 3 months | 1932 (86.7) | 2228 | 275 (90.2) | 305 | 527 (84.5) | 624 | 544 (85.8) | 634 | 586 (88.1) | 665 |
| Of those who reported using alcohol | ||||||||||
| Average drinking days per week [M (SD)] |
2.6 (2.0) |
1918 | 2.0 (1.5) | 273 | 2.5 (2.0) | 521 | 2.7 (2.0) | 541 | 2.8 (2.2) | 583 |
| One or more drinking days per week | 1725 (89.9) | 240 (87.9) | 463 (88.9) | 499 (92.2) | 523 (89.7) | |||||
| Three or more drinking days per week | 781 (40.7) | 84 (30.8) | 201 (38.6) | 240 (44.4) | 256 (43.9) | |||||
| Tobacco Cigarettes | ||||||||||
| Any use in past 3 months | 577 (25.8) | 2238 | 71 (23.2) | 306 | 199 (31.6) | 629 | 180 (28.1) | 641 | 127 (19.2) | 662 |
| Of those who reported using tobacco | ||||||||||
| ≥ 10 cigarettes per day | 279 (49.0) | 569 | 11 (15.7) | 70 | 93 (47.9) | 194 | 107 (59.8) | 179 | 68 (54.0) | 126 |
| Within 30 min of waking up | 376 (65.3) | 576 | 24 (33.8) | 71 | 137 (68.8) | 199 | 132 (73.7) | 179 | 83 (65.4) | 127 |
| Cannabis | ||||||||||
| Currently use cannabis product(s) at least once a month [N (%)] | 975 (45.1) | 2162 | 170 (56.1) | 303 | 319 (52.9) | 603 | 264 (42.9) | 616 | 222 (34.7) | 640 |
| Frequency of use [n (%)]b | ||||||||||
| Once a year or less frequently | 110 (9.4) | 1167 | 22 (10.5) | 209 | 36 (9.4) | 381 | 23 (7.4) | 311 | 29 (10.9) | 266 |
| A few times a year | 134 (11.5) | 26 (12.4) | 45 (11.8) | 34 (10.9) | 29 (10.9) | |||||
| A few times a month | 161 (13.8) | 35 (16.7) | 46 (12.1) | 41 (13.2) | 39 (14.7) | |||||
| 1–2 times a week | 124 (10.6) | 24 (11.5) | 41 (10.8) | 31 (10.0) | 28 (10.5) | |||||
| 3–6 times a week | 199 (17.1) | 35 (16.7) | 61 (16.0) | 54 (17.4) | 49 (18.4) | |||||
| Once a day or more frequently | 439 (37.6) | 67 (32.1) | 152 (39.9) | 128 (41.2) | 92 (34.6) | |||||
Note: All data are presented as n (%) unless otherwise noted; sample sizes vary, percentages calculated based on number of valid responses.
3.3. Latent class analysis results
Table 3 presents fit statistics for 2–5 class Latent Class Analysis (LCA) models. The four-class solution was selected based on the combination of theoretical interpretability, excellent entropy, and the lowest BIC; Fig. 3 displays the probability of endorsing each substance use frequency category by latent class. This solution consisted of 4 latent classes as follows: (1) “Primary Alcohol” class (approximately 57% of total sample) characterized by alcohol use of varying frequencies, predominantly cannabis abstinence or low frequency cannabis use, abstinence from “other” illicit drugs, and predominantly low likelihood of prescription misuse; (2) “Heavy Cannabis with Alcohol” (27%) characterized by high likelihood of heavy cannabis use (i.e., 16+ days) and varying frequencies of alcohol use; and (3) “Polysubstance Use” class (10%) characterized by the highest probabilities of endorsing frequent “other” illicit substance use. This class also has a wide range of estimated frequency of alcohol and cannabis use, in addition to “other” illicit substance use. Therefore, the “Polysubstance” use class likely contains significant variability and captures several substance use combinations, and lastly (4) “Prescription Medication Misuse with Alcohol” class (6%), characterized by moderate probabilities of endorsing moderate to high frequency prescription misuse, in addition to some alcohol use. Although it is of note that this class also had the highest likelihood of alcohol abstinence.
Table 3.
Latent class analysis fit statistics.
| # Classes | # Parameters | BIC | Log likelihood | Entropy | Size of Profiles (%) |
|---|---|---|---|---|---|
| 2 | 25 | 16,387.960 | −8097.243 | 0.50 | 58, 42 |
| 3 | 37 | 16,309.808 | −8007.865 | 0.73 | 52, 25, 23 |
| 4 | 51 | 16,309.683 | −7957.499 | 0.88 | 55, 30, 10, 6 |
| 5 | 64 | 16,353.168 | −7928.938 | 0.88 | 54, 29,6, 5, 4 |
Fig. 3.
Probability of endorsing each substance use frequency (none, mild, moderate, heavy) for each of the four latent classes: Primary Alcohol, Heavy Cannabis with some alcohol, polysubstance use, and prescription medication misuse with alcohol.
3.4. Multinomial logistic regression examining the association between age cohort and latent class membership
Table 4 presents the breakdown of participants' most likely latent class membership by age cohort. Odds of being in the “Primary Alcohol” cohort were higher (Odds Ratio = 1.86, 95% Confidence Intervals [1.54, 2.24], p < .001) for the 31–43 cohort, the 44–56 cohort (OR = 2.19 [1.81, 2.62], p < .001), and the 57–70 cohort (OR = 2.76 [2.32, 3.29], p < .001) compared to the 18–30 cohort. Compared to the youngest cohort, the odds of membership in the “Heavy Cannabis with Alcohol” class was not significantly different for the 31–43 cohort (OR = 0.98 [0.73, 1.32], p = .90) or the 44–56 cohort (OR = 0.75 [0.56, 1.02], p = .06) but was significantly lower in the 57+ cohort (OR = 0.44 [0.32, 0.60], p < .001). Similarly, compared to the 18–30 cohort, odds of membership in the “Polysubstance” class were not significantly different for the 31–43 cohort (OR = 1.53 [1.00, 2.35], p = .05) or the 44–56 cohort (OR = 0.78 [0.50, 1.23], p = .29), but were significantly lower in the 57+ cohort (OR = 0.37, [0.23, 0.61], p < .001). Odds of membership in the “Prescription Medication Misuse with Alcohol” class was not statistically different in the 31–43 cohort (OR = 2.30 [0.98, 5.35], p = .05) but was statistically higher in the 44–56 cohort (OR = 3.20 [1.42, 7.21], p = .01) and the 57+ cohort (OR = 2.53 [1.13, 5.71], p = .02) compared to the 18–30 cohort. Supplemental Tables 2 and 3 present odds ratios from multinomial models, using the 31–43 and 44–56 cohorts, respectively, as reference categories to obtain the remaining statistical comparisons between cohorts.
Table 4.
Latent class membership by age cohort.
| Class | Total sample | 18–30 | 31–43 | 44–56 | 57–70 |
|---|---|---|---|---|---|
| Alcohol Primary | 57% | 52% | 48% | 56% | 68% |
| Heavy Cannabis with Alcohol | 27% | 34% | 31% | 28% | 20% |
| Polysubstance | 10% | 11% | 16% | 9% | 5% |
| Rx. Misuse with Alcohol | 6% | 3% | 5% | 7% | 7% |
4. Discussion
As part of a larger project to identify correlates and potential risk factors for heavy substance use in women, this study aimed to elucidate substance use patterns in adult women in the community who reported past 30-day substance use and identify whether there are differences in use by age cohort. Results of latent class analyses demonstrated four main substance use patterns. The first two patterns focused on alcohol and cannabis use and represented the largest groups. More than half of the sample was in a class characterized predominantly by alcohol use, with a small likelihood of cannabis use, and just under one-third were in a class characterized by heavy cannabis use with some alcohol use. It is not surprising that the largest class had a predominant pattern primarily of alcohol, as alcohol is by far the most commonly used substance in the US (Substance Abuse and Mental Health Services Administration, 2024). Members of this group show a wide range of alcohol use frequency, from infrequent alcohol use of 1–5 days per month to more frequent use of 16 or more days per month. This was accompanied by a small likelihood (<20%) of endorsing any cannabis use. This contrasts with the group with predominantly heavy cannabis and some alcohol use. Women in this group are characterized by higher likelihood of heavy cannabis use of 16 or more days per month. Women in both these classes of use patterns had a low likelihood of prescription medication misuse and a very low likelihood of other illicit drug use.
Consistent with population base rates, fewer participants were in the polysubstance use class with heavy use of other illicit drugs, as well as some alcohol and cannabis use, and misuse of prescription medications; and a small minority of the sample belonged to a class with predominant misuse of prescription medications. Of note, the prescription medication misuse group was characterized by a high likelihood of frequent (16 or more days per month) prescription medication misuse, with a moderate likelihood of alcohol use. These two use patterns are understudied in women (McHugh et al., 2018) but pose risks for morbidity and mortality, including risks of accidents, overdose, chronic illness development, and greater co-occurring psychiatric disorders (Brady et al., 1993; Greenfield et al., 2007; Park et al., 2023; Rodriguez et al., 2022; White, 2022).
Significantly, the latent class patterns varied with age. The proportion of each age cohort in the primary alcohol use class increased with age cohort. This contrasts with the class where heavy cannabis use and some alcohol predominate, which is greatest in the younger age cohorts (i.e., 18–30, 31–43) and decreases in prevalence progressively with each older cohort. The two younger age cohorts also have the highest likelihood of other illicit drug use and membership in the polysubstance use class. By contrast, while the overall prevalence of the prescription medication misuse with some alcohol use class is low, it is most prevalent in the two oldest age cohorts. This may be reflective of greater exposure to prescribed medications, such as benzodiazepines and opioids in older age cohorts (Innes et al., 2024; Rhee, 2019). This study extends the current literature by demonstrating in a sample of community participants that these patterns of use are distinct and vary with age.
This study's findings that those in the older age cohorts are more likely to belong to the class with predominant use of alcohol is consistent with previous epidemiologic studies (Breslow et al., 2017; Han et al., 2017). For example, results from one study showed that between 1997 and 2014, binge drinking increased in people 60 years and older, and especially among women (Breslow et al., 2017). In addition, analyses of data from the National Survey on Drug Use and Health from 2005 to 2014 demonstrated that females 50 to 64 years had a significant 44.4% relative increase (6.3% to 9.1%) in binge drinking while males demonstrated a non-significant 9.7% relative increase (19.6% to 21.5%) in this same period (Han et al., 2017). Heavy alcohol use in older women carries multiple health risks including cognitive impairment, falls, liver disease, and breast cancer (Office of the Surgeon General (US), 2025).
As the landscape of cannabis legislation has continued to evolve along with societal norms, understanding cannabis use in women across the lifespan is of utmost importance. In our sample, it is the two younger age cohorts that have the highest prevalence of cannabis use (56.1% and 52.9% of 18–30- and 31–43-year-olds, respectively). Even so, it is notable that 45.1% of the overall sample endorsed using cannabis, including 34.7% of the oldest age cohort. Of note, recent national data demonstrate that females aged 18–30 exceeded males in cannabis use prevalence for the first time (Miech et al., 2025; National Institute on Drug Abuse, 2024). This study adds to the existing literature by demonstrating that cannabis use in adult women in a community sample varies with age cohorts including mean days of use per month. Understanding these patterns is essential to understanding the potential risks of alcohol, cannabis, and other drug use across women's adult lifespan.
Older age cohorts of women were also more likely to belong to the prescription medication misuse class, with 7% of each of the two older adult cohorts belonging to this class. While medication misuse is known to vary in the population by a range of factors including age, gender, and other factors (National Institute on Drug Abuse, 2020), the results of the present study extend findings of previous literature regarding substance use in older compared with younger women indicating a greater likelihood of prescription drug misuse combined with alcohol use in older women (Keyes, 2023; Schepis et al., 2022). Future research will be needed to parse differences among prescription drug types, as these vary across both age and gender (e.g., stimulants vs. benzodiazepines). By contrast, the two younger cohorts of women were more likely to use other illicit drugs and have polysubstance use across multiple substances compared with the two older age cohorts. Use of illicit drugs and polydrug use are reported in nationally representative epidemiologic data as being more common in adult men than women (Ellis et al., 2024; National Institute on Drug Abuse, 2020; Substance Abuse and Mental Health Services Administration, 2024). This survey of adult women demonstrates that in a community sample of women who used substances in the past month, illicit substance use co-occurring with use of multiple other substances may be most relevant among women in younger age cohorts.
Finally, while the U.S. has seen a decrease in tobacco smoking nationwide over the last five decades (Centers for Disease Control and Prevention, 2025), in this study 25.8% of adult women smoked combustible tobacco in the prior 3 months, with middle-aged women (31–43 and 44–56 years) having the highest prevalence (31.6% and 28.1%, respectively). Of those who smoked tobacco, 49.0% smoked more than 10 cigarettes per day, and 65.3% smoked within 30 min of awakening. These data indicate that in adult women in the community, nicotine and tobacco use remain significant and ought to be assessed and treated. Yet, use of nicotine replacement therapy and medication for nicotine cessation remains underutilized overall (Nielsen et al., 2024; Williams et al., 2025).
This study has several limitations. This is a cross-sectional, online survey designed to characterize the substance use patterns of a community-based population of adult women in the U.S. who endorsed at least one day of use of any substance in the past month. The study is not meant to replicate national epidemiological surveys such as the National Survey on Drug Use and Health. Because participants were recruited from a non-probability convenience sample drawn from the Sago panel, results should be interpreted with respect to this specific cohort and may not be generalizable to the entire population. However, this survey study extends results from previous studies by examining use patterns in a diverse sample of women in the community who endorse any use across multiple substances and age cohorts and finds that use of multiple substances is common, and patterns of use vary with age among adult women. Another limitation is that cross-sectional surveys capture only a single point in time, and causality cannot be inferred from association.
5. Conclusion
These limitations notwithstanding, this study's findings extend the literature by showing that in this diverse national community sample of adult women with past-month substance use, there is substantial heterogeneity in patterns of use of alcohol, cannabis, other illicit drug use, and prescription medication misuse. More than half used more than one substance, and use patterns varied with age cohort. Many of these groups are understudied (e.g., older women with prescription drug misuse; younger women with heavy cannabis use; older women with heavy alcohol use), highlighting the importance of research specifically targeting these high-risk subgroups. Heavy use of substances is a risk for developing substance use disorders and women are also at risk for an accelerated course of substance problems including medical and psychiatric consequences. Clinician screening for women's use of multiple substances through the adult lifespan would provide a significant opportunity for earlier detection and intervention for problematic substance use. Research is also critical to elucidate the risks for heavy substance use, including polysubstance use, through women's lifespan to design the most effective substance use prevention, early intervention, and treatment for adult women.
CRediT authorship contribution statement
Shelly F. Greenfield: Writing – original draft, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization. Allen J. Bailey: Writing – original draft, Formal analysis, Conceptualization. Laura C. Seidman: Writing – original draft, Project administration, Investigation, Data curation. Dawn E. Sugarman: Writing – review & editing. Kristin N. Javaras: Writing – review & editing. Laura A. Payne: Writing – review & editing. Victoria R. Votaw: Writing – review & editing. Nicole A. Barbaro: Writing – review & editing. R. Kathryn McHugh: Writing – review & editing.
Ethical approval
This study was conducted in accordance with the Declaration of Helsinki and approved by the Mass General Brigham Institutional Review Board (protocol 2024P001628, approved).
Ethics declaration
Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed.
This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. The Mass General Brigham Institutional Review Board granted an exemption. The authors provided the following explanation: The IRB determined that this project meets the criteria for exemption 45 CFR 46.104(d)(#). EXEMPTION (2) Research that only includes interactions involving educational tests (cognitive, diagnostic, aptitude, achievement), survey procedures, interview procedures, or observation of public behavior (including visual or auditory recording) if at least ONE of the following criteria is met: (i) The information obtained is recorded by the investigator in such a manner that the identity of the human subjects cannot readily be ascertained, directly or through identifiers linked to the subjects.
Declaration of generative AI and AI-assisted technologies in the writing process
The authors did not use AI in the writing of this manuscript.
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Shelly F Greenfield MD MPH reports financial support was provided by Harvard Radcliffe Institute for Advanced Study. Shelly Greenfield reports a relationship with National Institutes of Health that includes: funding grants. Shelly F Greenfield reports a relationship with Rutgers The State University of New Jersey that includes: speaking and lecture fees. Shelly F Greenfield MD MPH reports a relationship with Albert Einstein College of Medicine that includes: speaking and lecture fees. Dawn Sugarman reports a relationship with National Institutes of Health that includes: funding grants. Kate McHugh reports a relationship with National Institutes of Health that includes: funding grants. Kristin Javaras reports a relationship with National Institutes of Health that includes: funding grants. Laura A Payne reports a relationship with National Institutes of Health that includes: funding grants. Kristin Javaras reports a relationship with Broad Institute that includes: funding grants. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors would like to acknowledge the Dorothy Cherry and Timothy Morton Family Fund for Women and Addiction, McLean Hospital; Mary Beth and Chris Gordon Fellowship at the Harvard Radcliffe Institute (SFG); and the McLean Hospital Women's Mental Health Innovation Fund.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.abrep.2026.100748.
Appendix A. Supplementary data
Supplementary material 1
Supplementary material 2
Supplementary material 3
Data availability statement
The current approved institutional mechanism for data sharing is by individual data use agreements executed between interested parties.
References
- Bailey A.J., McHugh R.K. Why do we focus on the exception and not the rule? Examining the prevalence of mono- versus polysubstance use in the general population. Addiction. 2023;118(10):2026–2029. doi: 10.1111/add.16290. [DOI] [PubMed] [Google Scholar]
- Brady K.T., Grice D.E., Dustan L., Randall C. Gender differences in substance use disorders. The American Journal of Psychiatry. 1993;150(11):1707–1711. doi: 10.1176/ajp.150.11.1707. [DOI] [PubMed] [Google Scholar]
- Breslow R.A., Castle I.P., Chen C.M., Graubard B.I. Trends in alcohol consumption among older Americans: National Health Interview Surveys, 1997 to 2014. Alcoholism, Clinical and Experimental Research. 2017;41(5):976–986. doi: 10.1111/acer.13365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Centers for Disease Control and Prevention Adult Cigarette Smoking in the United States: Current and Historical Trends, Tobacco Use & Prevention. 2025. https://www.cdc.gov/tobacco/php/data-statistics/adult-data-cigarettes/index.html
- Ellis R.A., Bailey A.J., Jordan C., Shapiro H., Greenfield S.F., McHugh R.K. Gender differences in illicit drug access, use and use disorder: Analysis of National Survey on Drug Use and Health data. Journal of Psychiatric Research. 2024;175:118–122. doi: 10.1016/j.jpsychires.2024.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Greenfield S.F., Back S.E., Lawson K., Brady K.T. Substance abuse in women. The Psychiatric Clinics of North America. 2010;33(2):339–355. doi: 10.1016/j.psc.2010.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Greenfield S.F., Brooks A.J., Gordon S.M., Green C.A., Kropp F., McHugh R.K.…Miele G.M. Substance abuse treatment entry, retention, and outcome in women: A review of the literature. Drug and Alcohol Dependence. 2007;86(1):1–21. doi: 10.1016/j.drugalcdep.2006.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Grucza R.A., Bucholz K.K., Rice J.P., Bierut L.J. Secular trends in the lifetime prevalence of alcohol dependence in the United States: A re-evaluation. Alcoholism, Clinical and Experimental Research. 2008;32(5):763–770. doi: 10.1111/j.1530-0277.2008.00635.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Halloran E.C. Adult development and associated health risks. Journal of Patient-Centered Research and Reviews. 2024;11(1):63–67. doi: 10.17294/2330-0698.2050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Han B.H., Moore A.A., Sherman S., Keyes K.M., Palamar J.J. Demographic trends of binge alcohol use and alcohol use disorders among older adults in the United States, 2005-2014. Drug and Alcohol Dependence. 2017;170:198–207. doi: 10.1016/j.drugalcdep.2016.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harris P.A., Taylor R., Minor B.L., Elliott V., Fernandez M., O’Neal L.…Duda S.N. The REDCap consortium: Building an international community of software platform partners. Journal of Biomedical Informatics. 2019;95 doi: 10.1016/j.jbi.2019.103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harris P.A., Taylor R., Thielke R., Payne J., Gonzalez N., Conde J.G. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics. 2009;42(2):377–381. doi: 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Innes G.K., Ogden C.L., Crentsil V., Concato J., Fakhouri T.H. Prescription medication use among older adults in the US. JAMA Internal Medicine. 2024;184(9):1121–1123. doi: 10.1001/jamainternmed.2024.2781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kay A., Taylor T.E., Barthwell A.G., Wichelecki J., Leopold V. Substance use and women’s health. Journal of Addictive Diseases. 2010;29(2):139–163. doi: 10.1080/10550881003684640. [DOI] [PubMed] [Google Scholar]
- Keyes K.M. Vol. 107. Elsevier; 2023. Alcohol use in the older adult US population: Trends, causes, and consequences; pp. 28–31. [DOI] [PubMed] [Google Scholar]
- Keyes K.M., Grant B.F., Hasin D.S. Evidence for a closing gender gap in alcohol use, abuse, and dependence in the United States population. Drug and Alcohol Dependence. 2008;93(1–2):21–29. doi: 10.1016/j.drugalcdep.2007.08.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keyes K.M., Martins S.S., Blanco C., Hasin D.S. Telescoping and gender differences in alcohol dependence: New evidence from two national surveys. The American Journal of Psychiatry. 2010;167(8):969–976. doi: 10.1176/appi.ajp.2009.09081161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McHugh R.K., Nguyen M.D., Chartoff E.H., Sugarman D.E., Greenfield S.F. Gender differences in the prevalence of heroin and opioid analgesic misuse in the United States, 2015-2019. Drug and Alcohol Dependence. 2021;227 doi: 10.1016/j.drugalcdep.2021.108978. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McHugh R.K., Votaw V.R., Sugarman D.E., Greenfield S.F. Sex and gender differences in substance use disorders. Clinical Psychology Review. 2018;66:12–23. doi: 10.1016/j.cpr.2017.10.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McNeely J., Wu L.T., Subramaniam G., Sharma G., Cathers L.A., Svikis D.…Schwartz R.P. Performance of the tobacco, alcohol, prescription medication, and other substance use (TAPS) tool for substance use screening in primary care patients. Annals of Internal Medicine. 2016;165(10):690–699. doi: 10.7326/m16-0317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miech R.A., Johnston L.D., Patrick M.E., O'Malley P.M. Monitoring the future national survey results on drug use, 1975–2024: Overview and key findings for secondary school students. 2025. https://monitoringthefuture.org/results/annual-reports/
- Muthèn L.K., Muthèn B.O. Muthèn & Muthèn; 2017[1998]. Mplus User’s Guide. [Google Scholar]
- National Institute on Alcohol Abuse and Alcoholism . Alcohol’s effects on health. National Institutes of Health; 2025. Women and alcohol. [Google Scholar]
- National Institute on Drug Abuse . U.S. Department of Health and Human Services; 2020. Misuse of prescription drugs. [Google Scholar]
- National Institute on Drug Abuse . National Institute on Drug Abuse, National Institutes of Health; 2024. Cannabis and hallucinogen use among adults remained at historic highs in 2023. [Google Scholar]
- Nielsen E.M., Zhang J., Marsden J., Bays C., Moran W.P., Mauldin P.D.…Heincelman M. Is hospitalization a missed opportunity to intervene on tobacco cessation? The American Journal of the Medical Sciences. 2024;367(2):89–94. doi: 10.1016/j.amjms.2023.11.024. [DOI] [PubMed] [Google Scholar]
- Office of the Surgeon General (US) Alcohol and Cancer risk. 2025. https://www.hhs.gov/surgeongeneral/reports-and-publications/alcohol-cancer/index.html Washington, DC Retrieved from.
- Park Y., Dang E.P., Board A., Gilboa S.M., Ondersma S.J., Smid M.C.…Kim S.Y. Polysubstance use in pregnancy: Surveillance, interventions, and next steps. Journal of Women’s Health (2002) 2023;32(9):899–904. doi: 10.1089/jwh.2023.0341. [DOI] [PMC free article] [PubMed] [Google Scholar]
- R Core Team R: A Language and Environment for Statistical Computing. 2024. https://www.R-project.org/
- Rhee T.G. Coprescribing of benzodiazepines and opioids in older adults: Rates, correlates, and national trends. The Journals of Gerontology: Series A. 2019;74(12):1910–1915. doi: 10.1093/gerona/gly283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rockhill K.M., Black J.C., Iwanicki J., Abraham A. Polysubstance use profiles among the general adult population, United States, 2022. American Journal of Public Health. 2025;115(5):747–757. doi: 10.2105/ajph.2024.307979. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rodriguez A.S., Robinson L.D., Kelly P.J., Hudson S. Polysubstance use classes and health outcomes among women attending specialist substance use treatment services. Drug and Alcohol Review. 2022;41(2):488–500. doi: 10.1111/dar.13375. [DOI] [PubMed] [Google Scholar]
- Sago Sago survey platform. 2024. https://sago.com/en/solutions/quantitative/online-surveys/
- Schepis T.S., McCabe S.E., Ford J.A. Recent trends in prescription drug misuse in the United States by age, race/ethnicity, and sex. The American Journal on Addictions. 2022;31(5):396. doi: 10.1111/ajad.13289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Substance Abuse and Mental Health Services Administration . Center for Behavioral Health Statistics and Quality; 2021. Key substance use and mental health indicators in the United States: Results from the 2020 National Survey on Drug Use and Health. [Google Scholar]
- Substance Abuse and Mental Health Services Administration . Center for Behavioral Health Statistics and Quality; 2024. Key substance use and mental health indicators in the United States: Results from the 2023 National Survey on drug use and health. [Google Scholar]
- White J. Associations between illicit drug use in early adulthood and mortality: Findings from a National Birth Cohort. Preventive Medicine. 2022;159 doi: 10.1016/j.ypmed.2022.107058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wickham H. Springer-Verlag; 2016. ggplot2: Elegant graphics for data analysis.https://ggplot2.tidyverse.org [Google Scholar]
- Williams B.S., Fiore M.C., Hyland A., Slutske W.S. Vaping cessation methods used by young adults. JAMA Network Open. 2025;8(5) doi: 10.1001/jamanetworkopen.2025.12803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yee T.W. The VGAM package for categorical data analysis. Journal of Statistical Software. 2010;32(10):1–34. doi: 10.18637/jss.v032.i10. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary material 1
Supplementary material 2
Supplementary material 3
Data Availability Statement
The current approved institutional mechanism for data sharing is by individual data use agreements executed between interested parties.