Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Apr 1.
Published in final edited form as: Stress Health. 2022 Aug 26;39(2):361–371. doi: 10.1002/smi.3191

College Students’ Momentary Stress and Prescription Drug Misuse in Daily Life: Testing Direct Links and the Moderating Roles of Global Stress and Coping

Lauren M Papp 1, Chrystyna D Kouros 2, Laurie Armstrong 3, John J Curtin 4
PMCID: PMC9943780  NIHMSID: NIHMS1855366  PMID: 35994279

Abstract

Prior survey-based research has documented associations between greater levels of stress and increased prescription drug misuse behavior. These studies uniformly rely on assessments of both the stress experiences and the substance behavior after they occurred (commonly spanning 6- to 12-month retrospective timeframes). Less is known about the extent to which variations in momentary stress predict the actual occurrence of prescription misuse in daily life among college students with elevated risk for engaging in the behavior. In this study, 297 participants (69% females; Mage = 19.5 years, SDage = 0.71) completed a 28-day ecological momentary assessment (EMA) procedure that collected self-reported stress and other contextual experiences in moments preceding prescription drug misuse. Analyses tested the within-person association between momentary stress and prescription drug misuse and examined the extent to which the relation between stress and misuse was moderated by participants’ assigned sex or global stress and coping levels. Results from hierarchical generalized linear modeling indicated a significant within-person association between momentary stress (i.e., higher than usual relative to one’s own mean) and greater likelihood of prescription misuse in daily life, accounting for the number of stressors and timing covariates. No significant moderation by participant sex was found, and moderation effects by global stress and coping levels were not in the expected directions. Direct results highlight the role of momentary stress experiences on health-relevant substance behaviors and provide future directions for research and applied efforts.

Keywords: ecological momentary assessment, stress, mood, prescription drug misuse, young adults

BACKGROUND

A hallmark of college life for many young adults is the need to manage competing demands and respond to highly variable situations that occur throughout the day (Bouteyre et al., 2007; Russell et al., 2017), resulting in a period typified by increased stress levels with potentially wide-ranging effects on functioning (Sattler, 2019). Existing research documents associations between greater perceived stress (in both daily life and global levels) and increased occurrence of multiple forms of substance use (Magid et al., 2009; Metzger et al., 2017; Tavolacci et al., 2013). Among college students, these studies have emphasized the stress-related correlates of alcohol use (Perkins, 1999). As one example, Dvorak and colleagues (2018) conducted a daily process study involving multiple reports of college students’ mood states and drinking behavior across the day for 15 days. Their results found that ratings of current stress in daily life were lower after drinking had started (vs. before drinking) on drinking days as well as lower on drinking vs. non-drinking days (Dvorak et al., 2018). Prescription drug misuse – defined as a person taking a medication intended for someone else or taking their own medication for a different reason or in an alternative dose than intended – warrants increased attention as a behavioral response in this population (see Bonar et al., 2020; Robitalle & Collin, 2016). A small but growing area of research has identified college-based stress, especially pertaining to academics, as a robust correlate of stimulant medication misuse (e.g., Carlon et al., 2022; Norman & Ford, 2018; Sattler, 2019). However, studies in this area have been critiqued for their use of methodological approaches that rely on retrospective and global surveys to capture both stress and substance use behavior (Sattler, 2019; Schepis et al., 2021). Accordingly, the current study investigated the extent to which real-time variations in individuals’ moment-to-moment stress experiences directly predict their prescription misuse to establish and better understand these associations as they occur in everyday life.

The global links between elevated stress and prescription misuse have strong empirical support. Tam, Benotsch, and Weinstein (2020) studied college students’ prescription drug misuse behaviors and perceived stress in their lifetime via an online assessment of a large undergraduate sample (N=1052). Participants reported on misuse of 4 prescription medication classes (i.e., opioids, sedatives, anxiolytics, and stimulants) in their lifetime and in the past 3 months. Correlations indicated positive associations between perceived stress and misuse of all 4 medication classes. Similarly, Norman and Ford (2018) focused on young-adult college students’ academic strain (e.g., grade strain, academic challenges) as a predictor of prescription stimulant misuse “to help with academics” using a large (N=924) survey design. Results indicated consistent associations between academic strain indicators and stimulant misuse over the past academic year.

In a large-scale, representative study of 1,128 German adults, Bagusat et al. (2018) conducted cross-sectional surveys and structured interviews. The study assessed misuse of multiple types of psychoactive substances (e.g., cannabis, illicit drugs, and prescription medications taken not as prescribed). The assessment of substance use prompted for instances taken “…to improve your mental performance, improve your mood, relieve anxiety or nervousness, or manage stress.” Lifetime prevalence rates of prescription stimulants were lower compared to the other substance behaviors yet coping with stressful situations was the strongest predictor of this outcome. Higher reported stress levels over the past year predicted higher likelihood of (only) prescription stimulant misuse.

Drawing from stress and coping theories of substance use (Peck et al., 2018), another recent German study involving a large sample of students (N=2,203) randomly selected from four universities assessed misuse data for improving cognitive functioning in the past 6 months along with perceive stress in past month (Sattler, 2019). The medication misuse question prompted participants with examples of “better concentration, memory performance, or wakefulness.” The study hypothesized that misuse of stimulant medications may be utilized to cope with academic challenges in the college setting and further posited social support (indexed by number of social contacts) as a protective factor. Results revealed relatively low rates of the substance behavior; about 2% of participants endorsed misuse in the past 6 months. Higher perceived stress was directly associated with misuse occurrence. Moderation results were complex but generally suggested interaction effects not in the expected direction; as an example, among men with low perceived stress (but not among those with high stress), having more social support predicted greater likelihood of their prescription misuse.

Taken together, findings from these large surveys provide solid evidence for expecting associations between individuals’ stress and their prescription misuse behavior. However, additional study is warranted given the retrospective and global nature of the previous tests; that is, constructs were consistently collected (over months-long reporting periods) after the stress experience and the substance use had occurred. In addition, the approaches were commonly confounded, with the misuse assessments including reasons for the behavior (such as to reduce stress or improve academic performance) that were then tested in relation to participants’ global stress levels.

The Current Study

To overcome prior limitations, we introduce a more temporally precise assessment of stress experienced in the moment as a contextual correlate of prescription drug misuse in daily life. Ecological momentary assessment (EMA) offers a well-established approach to collecting near real-time data on stress and prescription misuse on multiple occasions throughout the day (Shiffman, 2009). Particular advantages include collecting hypothesized predictors of the behavior before it occurs, and thus before respondents have experienced any effect of the medication. In addition, the predictors are not linked to the behavior in this methodology, thereby improving upon studies that asked about the extent to which medication was used in order to deal with stress and therefore had confounded the motivations for the behavior and the misuse prior to conducting their statistical tests. Analysis of EMA data also accommodates moderating tests to identify characteristics of individuals that are reliably associated with variability (i.e., stronger or weaker) in within-person links between stress and prescription drug misuse.

We examined associations between stress intensity (over and above the number of stressors reported in the same moment) and prescription drug misuse in daily life among college students with elevated risk for engaging in the behavior. In line with models that posit reduction in stress via substance use (Bradford et al., 2013), we expected that more momentary stress than usual would be associated with greater likelihood of prescription misuse. We strengthened the rigor of the test by distinguishing stress intensity from the number of stressors experienced in the moment (Kaubrys et al., 2021) and accounting for other covariates (described in Analysis Plan). On the basis of prior work that has identified sex differences (Ford et al., 2014), we explored whether the within-person stress-misuse association was stronger for females as compared to males. We also assessed whether within-person links between momentary stress and prescription misuse were moderated by global stress and coping levels. Acknowledging mixed results observed in prior work (e.g., Sattler, 2019), we predicted that individuals reporting greater global stress (and lower global coping) would evidence stronger within-person links between their momentary stress intensity and likelihood of prescription drug misuse in daily life.

DATA AND METHODS

This study was drawn from an ongoing longitudinal project on college students’ daily substance use behaviors and health with a focus on prescription drug misuse (Papp et al., 2020). The University Institutional Review Board approved all procedures. Study enrollment and baseline participation took place between Fall 2017 and Fall 2019.

Participants and Enrollment Screening

Participants were recruited via flyers and announcements (e.g., newspaper ads, emails to enrolled students) that stated, “We are particularly interested in how people use prescription medications.” Prospective participants completed an online screening and the research team followed up with a telephone call to confirm eligibility. Inclusion criteria were being enrolled as a freshman or sophomore and being 18 to 21 years of age. An online screening questionnaire asked participants to think back over the past three months and indicate whether they used medications listed in any way a doctor did not intend, such as use without a prescription, increased amounts, more often, or longer than directed. The screener presented 4 prescription medication classes adapted from the 2015 National Survey on Drug Use and Health (Hughes et al., 2016): pain relievers, tranquilizers, stimulants, and sedatives or barbiturates, with common examples provided. Prospective participants responded to each medication question (Yes=1, No=0), and multiple classes could be endorsed.

Given the main objective of capturing prescription drug misuse in daily life during the original study, we oversampled for participants who endorsed recent prescription drug misuse. Of the 355 total participants enrolled, 300 (84.5%) endorsed misuse at screening (i.e., risk-group participants). Most (73.3%) endorsed recent misuse of one medication class at screening; the remainder endorsed misusing two or more classes. The majority of participants (77.3%) reported recent misuse of prescription stimulants; the next highest class was pain medications (22.3% of the sample), followed by prescription tranquilizers (19.3%), and prescription sedatives or barbiturates (11.3%).

The present analytic sample included 297 risk-group participants who returned ecological momentary assessment data. This analysis used reports completed on any reporting day; all follow-up reports were included. Sixty-nine percent of the sample was female, and the average age was 19.5 years (SD = 0.71). In terms of racial and ethnic background, 82.8% self-identified as White and 6.1% as Asian or Pacific Islander; 6.7% reported Hispanic or Latino/a background. Remaining participants self-identified as American Indian/Alaska Native (0.3%), Black or African American (1.7%), or reported multiple races (7.1%). Responses for another race (not specified) or unknown were 2%.

Procedure

Participants attended 2 lab sessions that were scheduled an average of 35 days apart and were trained to complete reporting procedures in daily life between the sessions. During the first session, participants completed informed consent procedures and survey measures (including demographic information) and were trained to use an iPod Touch application designed specifically for the present research. They chose a private password to access the application and completed a sample report in the lab. Access to all other device features was restricted. The scheduled reporting period started the following day. Although the typical reporting period was scheduled for the 28 days following the first lab session, some participants continued reporting until they returned their device at the second lab session; reports obtained across all days were retained in the current analyses to maximize statistical power. In the second lab visit, participants returned their devices, completed a Timeline Follow-back Interview, provided feedback on their daily life reporting experience, and received compensation (details in Papp et al., 2020).

Both signal-contingent and event-based assessments were administered. Signal-contingent reporting involved responding to a device prompt sent at random times within four windows (8–11:30am, 11:30am-3pm, 3–7pm, 7–11pm). Participants were instructed to respond as soon as possible. Participants were also trained to self-initiate a report any time they intended to take one of four medications in any way a doctor did not direct them to use it (i.e., event-based assessment). There was no limit on the timing or number of event-based reports that could be completed. To reduce burden, a signal-contingent prompt was not sent within 2 hours after a self-initiated, event-based report had been completed. EMA report questions included intentions to misuse and focused on potential triggers of prescription drug misuse (e.g., mood states, pain, stressors, other substance behaviors). Signal-contingent and event-based assessments were identical. Therefore, report type was indistinguishable in the resulting data files. If misuse intention of one or more of the medication classes was endorsed in the EMA report, participants were then sent a brief follow-up report 15 minutes later (Thrul et al., 2014) to assess misuse behaviors that occurred since the completion of the associated report; participants were instructed to respond to the follow-up within 15 minutes. As Papp and colleagues (2020) reported, our technology at the time of the study allowed participants to complete the follow-up outside of the requested response window. The majority of follow-ups (75%) were completed within an hour of the associated EMA report. All follow-up reports obtained were retained to maximize statistical power.

Measured Variables

Momentary Outcome: Prescription Drug Misuse

Participants responded (Yes=1, No=0) on signal- and event-contingent reporting to the question, “Are you about to take a medication listed here, in any way a doctor did not direct you to use it?” for 4 classes of medication (sedatives or sleeping pills, tranquilizers or anxiety medications, stimulants, and pain relievers). When people endorsed the intention question for one or more medications, they received a brief follow-up prompt 15 minutes after the associated report was completed. Participants then responded (Yes=1, No=0) to the question, “Have you recently taken a medication listed here not as prescribed?” for the 4 classes of medication. Misuse behavior of any of the four prescription medication classes is the primary outcome of interest; the outcome variable was scored 1 when the behavior was endorsed on the follow-up and 0 when the behavior was not endorsed on the follow-up or when a follow-up was not administered or completed. In instances when the intention questions were left blank (occurred on 57 EMA reports; described under Preliminary Results), the misuse behavior variable was left as missing data.

Momentary Predictor: Stress

Participants indicated whether they were currently experiencing 8 common events or situations in daily life (e.g., academic, romantic, work, financial). The items were presented as a hassles checklist, based on the work of Bolger and Schilling (1991) and adapted to college life by including a category for academic events (Lester, 2014) and broadening “spouse” language to romantic events. When endorsed, participants were prompted to rate the event on a scale from −3 (very unpleasant) to 3 (very pleasant); 0 reflected a neutral event. Following Jacobs et al. (2007), scores of −3, −2, or −1 were recoded to 3, 2, 1, respectively, and scores of 0 to 3 (i.e., ratings of neutral or pleasant events) were recoded to 0. When multiple unpleasant events were indicated in a given moment, the number of events (i.e., stressors) was summed and the ratings were summed to reflect total stress intensity experienced in the moment. When the moment included no unpleasant event, the moment’s stress rating and number of stressors were coded as 0.

Moderators

Participants self-reported their sex (as Male or Female) on a demographic survey completed during the first lab visit, conducted prior to the EMA reporting period. Participants also reported on their global stress and coping abilities (independent of specific stress events) using the 12-item Rhode Island Stress and Coping Inventory (RISCI; Fava, Ruggiero, & Grimley, 1998). Participants rated the frequency of items (e.g., “I was pressured by others” and “I was able to cope with unexpected problems”) during the past month from 1 (Never) to 5 (Frequently). Item scores were summed to obtain stress (7 items) and coping (5 items) scales. Sound psychometric properties have been demonstrated in prior work, including evidence of good reliability and validity during instrument development (Fava et al., 1998) and good scale reliability in research involving college students (Wang & Castañeda-Sound, 2008). The scales showed good reliability in the present sample (stress scale: Cronbach’s α = .85; coping scale: Cronbach’s α = .86).

Covariates

Potential effects of reporting across time were accounted for by including a moment indicator (e.g., 1–4, reflecting the report number on that day) for each level 1 report. A moment-level covariate reflecting a weekend indicator (Friday afternoon through Sunday morning =1, Sunday afternoon through Friday morning =0; see Kouros et al., 2021) was also included at level 1.

Analytic Plan

As preliminary analyses, we examined descriptive statistics and tested correlations between the person-level study variables. For the main research questions, we used multilevel modeling to examine within-person associations between momentary stress and prescription misuse behavior in daily life, and to test the role of the moderators. Models were run as a three-level hierarchical generalized linear model (HGLM) with Bernoulli distribution to account for the nested data (i.e., moments nested within days, days nested within people) and the binary dependent variable. Models were tested using HLM v.8 (Raudenbush et al., 2019). This analytic approach has been used successfully in prior studies with infrequent substance use behaviors (e.g., Scott et al., 2018) and the current study’s models successfully converged with no errors or warnings. Variance components were tested as random effects and fixed effect results from the population-average model with robust standard errors are presented. We followed centering guidelines for disentangling within-person from between-person associations (Bolger & Laurenceau, 2013). Stress ratings at level 1 were person-centered to represent the extent to which one’s momentary stress experience was higher or lower than their average stress rating across the reporting period. Although intensity of stress experienced in the moment was our focal predictor of misuse, we also person-centered number of stressors experienced. The person-mean values of the momentary stress variables were included as covariates at level 3. Thus, analyses provided a rigorous test of the extent to which within-person variation in stress intensity in the moment predicted occurrence of prescription drug misuse in the same moment, accounting for momentary variation in number of stressors. Analyses further tested between-person associations, or the extent to which individuals who experienced higher stress intensity ratings and more instances of stressors, on average, also reported more misuse behavior across the reporting period. To rule out any potential time effects, we followed multilevel guidelines (Bolger & Laurenceau, 2013) and accounted for report number across the day (level 1 covariate). Prior research has shown differences in college students’ substance use behaviors (typically alcohol consumption; Maggs et al., 2011) across weekend and weekdays. Accordingly, we included an indicator of whether the report occurred during the weekend or not (level 1 covariate).

To test whether participant sex and global stress and coping moderated the within-person link between momentary stress and prescription drug misuse, these factors were entered (grand-mean centered) at level 3 as cross-level moderators of the slopes for the level 1 momentary stress predictor. We used Preacher et al.’s (2006) online tool to probe the effect of significant moderators at low, moderate, and high levels; we chose to plot these respective values at the 16th, 50th, and 84th percentiles of the moderator, as recommended by Hayes (2013). The moderators were also included as level 3 intercept controls, along with the between-person variables and timing covariates described above. Each moderator was tested separately.

We converted adjusted odds ratios to Cohen’s d values (Chinn, 2000) to interpret the relative magnitude of statistically significant relationships (p < .05) across predictor variables and across models. The analysis plan was preregistered at Open Science Framework (osf.io/wuydc/).

RESULTS

Preliminary Results

Papp et al. (2020) provides additional information on the broader study’s protocol for reporting in daily life, evidence of participant EMA compliance, and the reported prescription misuse behaviors by medication class. The EMA reports in level 1 included minimal missing data; stress ratings and prescription misuse intention/behavior variables were missing on ≤0.2% of EMA reports. These were left as missing values and handled by HLM. There was no missing data that affected the scoring of the global stress and coping measure. Thus, no statistical adjustments for missingness were required in level 3.

Participants returned an average of 80.45 momentary reports (SD = 25.85, range: 11–130). The majority of participants (n = 278, 93.6%) reported experiencing at least one stressful event over the EMA period; those experiencing stress reported an average of 15.65 moments (SD = 15.02, range = 1–97) during the reporting period. The prescription misuse outcomes were less common. Across the EMA period, 146 (49.2%) participants reported intending to misuse a prescription medication at least once and 105 (35.4%) endorsed actual misuse behavior at least once in daily life. Among those endorsing any prescription drug misuse intention or behavior, participants reported intention an average of 4.25 (SD = 4.07, range = 1–25) times and actual behavior an average of 3.09 (SD = 3.12, range = 1–17) times during the reporting period. It is worth noting that nearly half (297/621 = 47.8%) of the EMA reports indicating any prescription drug misuse intention (completed by 110 of the participants) did not result in actual misuse behavior, due to either no follow-up completed (for 150 of the intention reports) or no misuse behavior reported on the follow-up (for the remaining 147 intention reports).

As shown in Table 1, male participants reported lower occurrence of momentary stress in daily life and had lower global stress levels. Aggregating across the reporting period, momentary stress occurrence was positively correlated with prescription drug misuse in daily life (both intention and behavior) and with higher general stress levels, on the one hand, and with lower general coping levels, on the other hand. Levels of prescription drug misuse intention and behavior in daily life were highly intercorrelated. Levels of prescription misuse intention and behavior in daily life were similarly positively associated with global stress levels but were not reliably related to general coping levels. Participants’ global stress and coping levels were inversely related to each other.

Table 1.

Descriptive Statistics and Correlations for Person-level Study Variables

M (or n, %) SD Range 1 2 3 4 5 6 7
1. Male 92, 31% 0–1 --
2. Stress events in daily lifea 0.19 0.18 0–.99 −.17** --
3. Stress intensity in daily lifeb 0.46 0.92 .01–12.5 −.12* .75** --
4. Prescription drug misuse in daily lifec (primary outcome) 0.01 0.03 0–.18 −.11 .21** .10 --
5. Prescription drug misuse intention in daily lifec (secondary outcome) 0.03 0.05 0–.30 −.07 .20** .11 .80** --
6. RISCI stress subscale 21.30 5.49 9–35 −.12* .29** .23** .14* .16** --
7. RISCI coping subscale 18.88 3.30 9–25 .06 −.23** −.23** −.03 −.09 −.53** --
Open in a new tab

Note. N = 297 participants.

a

Calculated as the number of moments with one or more stress events / n momentary reports for each participant.

b

Calculated as the average momentary stress intensity rating among 278 participants who reported one or more stress events.

c

Calculated as the sum of instances of prescription drug misuse intention or behavior / n momentary reports for each participant.

*

p < .05.

**

p < .01.

Central Results: Predicting Prescription Drug Misuse Behavior

Central hypotheses were tested in a series of multilevel models that included momentary stress intensity ratings and number of stressors as predictors of prescription misuse in daily life (described under Analytic Plan).

The direct association between momentary stress and prescription drug misuse is shown in Table 2 (Model 1). Results from multilevel modeling confirmed our central hypothesis and showed that participants were more likely to engage in prescription misuse in daily life in moments of their higher-than-usual stress (i.e., a positive within-person result), accounting for number of stressors they experienced in the moment. Between-person effects indicated no reliable links between individuals’ average stress ratings or average number of stress events (i.e., averaged across EMA reports) and their likelihood of engaging in prescription misuse.

Table 2.

Associations between Momentary Stress and Prescription Drug Misuse in Daily Life: Direct and Moderated Results

Fixed effects predictors AOR p-value 95% CI d-value
Model 1: Direct Association
 Intercept 0.078 <.001 [0.076, 0.080] --
 WP stress rating 1.084 <.001 [1.057, 1.112] 0.04
 BP stress rating 0.881 0.186 [0.729, 1.064] --
 WP number of stress events 1.075 0.004 [1.024, 1.128] 0.04
 BP number of stress events 1.634 0.020 [1.082, 2.468] 0.27
Model 2: Moderated by Sex
 Intercept 0.075 <.001 [0.074, 0.077] --
  X Sex 0.895 <.001 [0.861, 0.931] −0.06
 WP stress rating 1.087 <.001 [1.060, 1.115] 0.05
  X Sex 1.013 0.289 [0.989, 1.037] --
 BP stress rating 0.889 0.211 [1.053, 2.376] --
 WP number of stress events 1.076 0.003 [1.025, 1.130] 0.04
 BP number of stress events 1.582 0.027 [1.053, 2.376] 0.25
Model 3: Moderated by Global Stress
 Intercept 0.076 <.001 [0.075, 0.078] --
  X Global stress 1.013 <.001 [1.010, 1.017] 0.07
 WP stress rating 1.088 <.001 [1.060, 1.117] 0.05
  X Global stress 0.997 0.008 [0.994, 0.999] −0.002
 BP stress rating 0.833 0.039 [0.701, 0.990] −0.101
 WP number of stress events 1.080 0.002 [1.028, 1.135] 0.04
 BP number of stress events 1.807 0.002 [1.249, 2.615] 0.33
Model 4: Moderated by Global Coping
 Intercept 0.078 <.001 [0.077, 0.080] --
  X Global coping 0.996 0.125 [0.991, 1.001] --
 WP stress rating 1.085 <.001 [1.057, 1.114] 0.04
  X Global coping 1.007 <.001 [1.004, 1.010] 0.004
 BP stress rating 0.875 0.165 [0.725, 1.057] --
 WP number of stress events 1.067 0.009 [1.016, 1.121] 0.04
 BP number of stress events 1.639 0.018 [1.087, 2.469] 0.27
Open in a new tab

Note. N = 297 participants. Results from population-average models with robust standard errors presented. Sex coded as female = −0.5 and male = 0.5. AOR = adjusted odds ratio. WP = within-person. BP = between-person. Models account for momentary report number across the day and a weekend indicator. Following Chinn (2000), d-values were calculated to facilitate interpretation of statistically significant (p < .05) fixed effect associations.

Participant sex was tested as a moderator of the within-person association between momentary stress and prescription misuse in daily life (Table 2, Model 2). No reliable support for moderation was found, and participant sex was not considered further.

Next, two models tested moderators of the within-person association between stress and prescription misuse by participants’ global stress and coping levels. The model for global stress (Table 2, Model 3) indicated reliable support for moderation. Simple slope analyses were conducted by plotting regression lines at relatively low, moderate, and high levels of the global stress scale (see Figure 1a). Findings from these tests revealed positive within-person links between momentary stress intensity and prescription drug misuse at low (b = 0.102, SE = 0.016, p < .001), moderate (b = 0.086, SE = 0.014, p < .001), and high (b = 0.066, SE = 0.014, p <. 001) levels of global stress. Effects were thus stronger at lower levels of global stress, which was not consistent with the hypothesized direction.

graphic file with name nihms-1855366-f0001.jpg

Open in a new tab

Figures show the moderating roles of global stress (Figure 1a) and global coping (Figure 1b) on the within-person association between momentary stress intensity and prescription misuse in daily life. The estimated log-odds of prescription misuse behavior (i.e., the dependent variable) were converted to probabilities for ease of interpretation, using the following equation: exp(log-odds value)/ [1 + exp(log-odds value)].

Testing the model with global coping (Table 2, Model 4) also indicated reliable evidence of moderation. Results from the simple slope tests (shown in Figure 1b) indicated positive within-person links between momentary stress and prescription drug misuse at low (b = 0.060, SE = 0.013, p < .001), moderate (b = 0.075, SE = 0.013, p < .001), and high (b = 0.104, SE = 0.016, p <. 001) levels of global coping. Thus, the effects were stronger at higher levels of global coping, which again was not consistent with the expected direction.

Across all models, the positive within-person association between momentary stress intensity and prescription drug misuse remained significant; effect sizes of the reliable findings from direct and moderation tests were in the very small to small range (see d-values in Table 2).

Secondary Result: Predicting Intention to Misuse Prescription Drugs

Acknowledging the discrepancy between the number of EMA reports that indicated participants’ intention to misuse prescription drugs and those that captured actual misuse behavior, we conducted an additional direct test between momentary stress and the outcome of misuse intention in daily life. Similar to the central test (reported above and in Table 2, Model 1), the secondary model accounted for number of stressors experienced in the moment and included the same between-person predictors and timing covariates. Results showed a significant direct link between individuals’ higher-than-average stress ratings (i.e., a positive within-person result) and their intention to misuse prescription drugs in daily life (AOR = 1.09, 95% CI [1.047, 1.131], p < .001, d = 0.05). Comparing results obtained from the central and secondary model tests indicated positive links between college students’ experiencing elevated stress in the moment and their increased likelihood of prescription misuse intention and behavior in daily life; across outcomes, the within-person effect sizes were consistently very low.

DISCUSSION

This study found that college students were more likely to engage in prescription drug misuse in daily life during moments they evaluated as higher stress than typical for them. We utilized a real-time reporting approach that captured participants’ ratings of stress intensity (controlling for the number of stress events experienced) prior to the occurrence of misuse behavior. Moreover, our statistical modeling approach separated both the evaluations of stress intensity and instances of stress into between- and within-person predictors and analyses accounted for multiple covariates. Within-person variation in stress remained a robust predictor of misuse behavior in daily life across direct and moderating tests. The results align with the growing body of results based on cross-sectional findings that link person-level experiences of stress to greater prescription misuse behavior.

Results from hypothesized moderation tests portrayed a less consistent picture. In particular, the direct, within-person association was not moderated by participant sex. Ford et al. (2014) previously argued that females display greater risk for prescription misuse behavior and for engaging in misuse in response to strain. Our person-level findings did not indicate reliable associations between participant sex and extent of prescription misuse reported in daily life; further, sex did not moderate the positive association between momentary stress and misuse in daily life. It is possible that our approach of combining across multiple stressors (e.g., interpersonal, academic, work) in daily life masked potential interactions between participant sex and stressful experiences in predicting the likelihood of misuse.

Additionally, moderation tests involving global stress and coping levels resulted in findings that were not consistent with our hypotheses. The positive within-person link was weaker among individuals with higher levels of global stress (and stronger among those with lower levels of global coping). Interestingly, Sattler (2019) considered similar constructs in a relevant sample and also found that global support levels did not moderate in the expected directions. Sattler found that among men with low stress, having more social supports (tested as a protective factor) increased prescription stimulant misuse; thus, it is possible that having greater social connections facilitates access to diverted prescription medications, in particular, and to increased deviant behavior, more broadly (Sattler, 2019). Another hypothesis for future research to consider is that individuals at risk for engaging in prescription misuse who have less global stress might react more strongly when they experience heightened momentary stress in daily life; however, we acknowledge this pattern differs from previous findings based on individuals not sampled according to substance use behaviors (Stawski et al., 2008). Together, results from this line of study indicate that prescription misuse functions differently than other common college-based substance behaviors wherein greater “risk” factors are associated with greater use of the substance, such as alcohol. Further, it is possible that prescription misuse serves as a short-term coping strategy although longer-term implications of the behavior are less established. We plan to conduct follow-up tests with longitudinal assessments that are underway to explore these possibilities.

In terms of practical implications, the current findings indicate that targeting in the moment hassles and stressors may be more fruitful than general stress tendencies to reduce prescription misuse behavior. Descriptively, instances of stress (defined as here as momentary events that were rated to some degree as unpleasant) were common in the sample. For college students with elevated risk for engaging in misuse behavior (due to prior misuse, access to diverted prescriptions, or their own prescription for medical reasons), raising awareness that these unpleasant experiences directly precede prescription misuse could help inform college-based prevention and intervention efforts. Moore and colleagues (2014) utilized an innovative wastewater analysis to establish that higher amounts of prescription psychostimulant medications were present in a college campus wastewater at midterms and finals compared to at the start of semester. Thus, the present findings can be used with other methods to anticipate points in the academic calendar when students might be experiencing higher-than-usual stress experiences.

Limitations and Future Directions

Several limitations warrant consideration. First, instances of prescription misuse in daily life were relatively low (Papp et al., 2020), despite our reliance on a sample with elevated likelihood of the behavior. It is possible that participants underreported their substance use. Additional work is needed to validate EMA assessments of prescription misuse behavior to strengthen confidence in self-reported usage trends obtained in daily life. We have reported the full extent of our recruitment efforts and numerous suggestions for increasing capture of the behavior in future studies, including offering the EMA reporting procedure on participants’ own devices (Papp et al., 2020). Also, the reporting procedures were based on capturing instances of misuse behavior rather than stressful events. Thus, stress experiences that occurred out of the reporting windows were not represented. Relatedly, because participants were asked to initiate EMA reports when they intended to engage in prescription drug misuse, the study does not address how often participants experienced high levels of stress and used a different coping technique.

Second, given the relatively low misuse frequencies, we combined across indicators of misuse of multiple prescription classes. This approach aligns with other studies in the field and provides a comprehensive view of the behavior in daily life. Consistent with US-based representative samples (Schepis et al., 2018), stimulant misuse was the most frequently reported medication class in our sample and likely drove the pattern of findings. However, recent work identifies differential correlates of college students’ misuse of prescription stimulants vs. opioid medications (Panlilio et al., 2021) and thus encourages continued examination of separate classes as datasets allow.

Third, our sample was relatively homogeneous and included an overrepresentation of White participants and females. Our study does not address stress contexts that have been associated with young adult substance use in prior work, including minoritized identities such as LGBTQ (Felner et al., 2020) and racial and ethnic (Sumstine et al., 2018) statuses. Thus, the generalizability of our findings is limited.

Despite these limitations, the current study provides avenues for future research and applied efforts. Prevention and intervention efforts aimed at promoting healthy responses to stress on college campuses could raise awareness of prescription misuse behavior patterns and consequences. As noted, future research should examine the extent of agreement in college students’ reports of misuse behavior collected across multiple methods, similar to work based on other salient substances in this population (e.g., alcohol, Patterson et al., 2019). Managing stress is a fundamental goal in the daily lives of young adult college students and can present challenges to maintaining longer-term physical and psychological health (Tessner et al., 2011; van der Steen et al., 2017). Given the established role of stress in maintaining substance use over time, it remains critical to advance efforts to understand and prevent prescription misuse in young adults’ daily lives.

Acknowledgments

Research reported in this publication was supported by the National Institute on Drug Abuse of the National Institutes of Health under Award Number R01DA042093. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Requests for access to data supporting the findings of this study can be directed to the corresponding author. We gratefully acknowledge the UWellness Study’s participants and research assistants. We have no known conflict of interest to disclose.

Contributor Information

Lauren M. Papp, University of Wisconsin-Madison

Chrystyna D. Kouros, Southern Methodist University

Laurie Armstrong, University of the Virgin Islands.

John J. Curtin, University of Wisconsin-Madison

References

  1. Bagusat C, Kunzler A, Schlecht J, Franke AG, Chmitorz A, & Lieb K (2018). Pharmacological neuroenhancement and the ability to recover from stress – a representative cross-sectional survey among the German population. Substance Abuse Treatment, Prevention, and Policy, 13(37). 10.1186/s13011-018-0174-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bolger N, & Laurenceau J-P (2013). Intensive longitudinal methods: An introduction to diary and experience sampling research. Guilford Press. [Google Scholar]
  3. Bolger N, & Schilling EA (1991). Personality and the problems of everyday life: The role of neuroticism in exposure and reactivity to daily stressors. Journal of Personality, 59, 355–386. 10.1111/j.1467-6494.1991.tb00253.x [DOI] [PubMed] [Google Scholar]
  4. Bonar EE, Coughlin L, Roche JS, Philyaw-Kotov ML, Bixler EA, Sinelnikov S, Kolosh A, Cihak MJ, Cunningham RM, & Walton MA (2020). Prescription opioid misuse among adolescents and emerging adults in the United States: A scoping review. Preventive Medicine, 132, 105972. 10.1016/j.ypmed.2019.105972 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bouteyre E, Maurel M, & Bernard J-L (2007). Daily hassles and depressive symptoms among first year psychology students in France: The role of coping and social support. Stress & Health, 23(2), 93–99. 10.1002/smi.1125 [DOI] [Google Scholar]
  6. Bradford DE, Shapiro BL, & Curtin JJ (2013). How bad could it be? Alcohol dampens stress responses to threat of uncertain intensity. Psychological Science, 24(12), 2541–2549. 10.1177/0956797613499923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carlon HA, Peters G, & Villarosa-Hurlocker MC (2022). When stimulant use becomes problematic: Examining the role of coping styles. Substance Use & Misuse, 57(3), 442–451. 10.1080/10826084.2021.2019774 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chinn S (2000). A simple method for converting an odds ratio to effect size for use in meta-analysis. Statistics in Medicine, 19, 3127–3131. 10.1002/1097-0258(20001130)19:22<3127::AID-SIM784>3.0.CO;2-M [DOI] [PubMed] [Google Scholar]
  9. Dvorak RD, Stevenson BL, Kilwein TM, Sargent EM, Dunn ME, Leary AV, & Kramer MP (2018). Tension reduction and affect regulation: An examination of mood indices on drinking and non-drinking days among university student drinkers. Experimental and Clinical Psychopharmacology, 26(4), 377–390. 10.1037/pha0000210 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fava JL, Ruggiero L, & Grimley DM (1998). The development and structural confirmation of the Rhode Island Stress and Coping Inventory. Journal of Behavioral Medicine, 21, 601–611. 10.1023/A:1018752813896 [DOI] [PubMed] [Google Scholar]
  11. Felner JK, Wisdom JP, Williams T, Katuska L, Haley SJ, Jun H-J, & Corliss HL (2020). Stress, coping, and context: Examining substance use among LGBTQ young adults with probable substance use disorders. Psychiatric Services, 71, 112–120. https://doi.org/appi.ps.201900029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Ford JA, Reckdenwald A, & Marquardt B (2014). Prescription drug misuse and gender. Substance Use & Misuse, 49, 842–851. 10.3109/10826084.2014.880723 [DOI] [PubMed] [Google Scholar]
  13. Hayes AF (2013). Mediation, moderation, and conditional process analysis: A regression-based approach. Guildford Press. [Google Scholar]
  14. Hughes A, Williams MR, Lipari RN, Bose J, Copello EAP, & Kroutil LA (2016). Prescription drug use and misuse in the United States: Results from the 2015 National Survey on Drug Use and Health. NSDUH Data Review. September 2016. Accessed at www.samhsa.gov/data/sites/default/files/NSDUH-FFR2-2015/NSDUH-FFR2-2015.htm [Google Scholar]
  15. Jacobs N, Myin-Germeys I, Derom C, Delespaul P, van Os J, & Nicolson NA (2007). A momentary assessment study of the relationship between affective and adrenocortical stress responses in daily life. Biological Psychology, 74(1), 60–66. 10.1016/j.biopsycho.2006.07.002 [DOI] [PubMed] [Google Scholar]
  16. Kaubrys M, Baker MR, Frazier PA, & Nguyen-Feng VN (2021). Relations among daily stressors, childhood maltreatment, and sleep in college students. Journal of Counseling Psychology, 68(4), 489–500. 10.1037/cou0000549 [DOI] [PubMed] [Google Scholar]
  17. Kouros CD, Papp LM, Kelly BC, & Blumenstock SM (2021). College-based social and situational predictors of real-time prescription drug misuse in daily life. Drug and Alcohol Dependence, 227, 108969. 10.1016/j.drugalcdep.2021.108969 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lester D (2014). College student stressors, depression, and suicidal ideation. Psychological Reports, 114(1), 293–296. 10.2466/12.02.PR0.114k10w7 [DOI] [PubMed] [Google Scholar]
  19. Maggs JL, Williams LR, & Lee CM (2011). Ups and downs of alcohol use among first-year college students: Number of drinks, heavy drinking, and stumble and pass out drinking days. Addictive Behaviors, 36(3), 197–202. 10.1016/j.addbeh.2010.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Magid V, Colder CR, Stroud LR, Nichter M, Nichter M, & TERN Members. (2009). Negative affect, stress, and smoking in college students: Unique associations independent of alcohol and marijuana use. Addictive Behaviors, 34(11), 973–975. 10.1016/j.addbeh.2009.05.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Metzger IW, Blevins C, Calhoun CD, Ritchwood TD, Gilmore AK, Stewart R, & Bountress KE (2017). An examination of the impact of maladaptive coping on the association between stressor type and alcohol use in college. Journal of American College Health, 65(8), 534–541. 10.1080/07448481.2017.1351445 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Moore DR, Burgard DA, Larson RG, & Ferm M (2014). Psychostimulant use among college students during periods of high and low stress: An interdisciplinary approach utilizing both self-report and unobtrusive chemical sample data. Addictive Behaviors, 39, 987–993. 10.1016/j.addbeh.2014.01.021 [DOI] [PubMed] [Google Scholar]
  23. Norman L, & Ford J (2018). Undergraduate prescription stimulant misuse: The impact of academic strain. Substance Use & Misuse, 53(9), 1482–1491. 10.1080/10826084.2017.1413393 [DOI] [PubMed] [Google Scholar]
  24. Panlilio LV, Stull SW, Bertz JW, Burgess-Hull AJ, Lanza ST, Curtis BL, Phillips KA, Epstein DH, & Preston KL (2021). Beyond abstinence and relapse II: Momentary relationships between stress, craving, and lapse within clusters of patients with similar patterns of drug use. Psychopharmacology, 238, 1513–1529. 10.1007/s00213-021-05782-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Papp LM, Barringer A, Blumenstock SM, Gu P, Blaydes M, Lam J, & Kouros CD (2020). Development and acceptability of a method to investigate prescription drug misuse in daily life: An ecological momentary assessment study. JMIR mHealth and uHealth, 8(10), e21676. 10.2196/21676 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Patterson C, Hogan L, & Cox M (2019). A comparison between two retrospective alcohol consumption measures and the daily drinking diary method with university students. The American Journal of Drug and Alcohol Abuse, 45(3), 248–253. 10.1080/00952990.2018.1514617 [DOI] [PubMed] [Google Scholar]
  27. Peck JH, Childs KK, Jennings WG, & Brady CM (2018). General strain theory, depression, and substance use: Results from a nationally representative, longitudinal sample of White, African-American, and Hispanic adolescents and young adults. Journal of Child & Adolescent Substance Abuse, 27(1), 11–28. 10.1080/1067828X.2017.1396516 [DOI] [Google Scholar]
  28. Perkins HW (1999). Stress-motivated drinking in collegiate and postcollegiate young adulthood: Life course and gender patterns. Journal of Studies on Alcohol, 60(2), 219–227. 10.15288/jsa.1999.60.219 [DOI] [PubMed] [Google Scholar]
  29. Preacher KJ, Curran PJ, & Bauer DJ (2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31(4), 437–448. 10.3102/10769986031004437 [DOI] [Google Scholar]
  30. Raudenbush SW, Bryk AS, Cheong YF, & Congdon R (2019). HLM 8 for Windows [Computer software]. Scientific Software International, Inc. [Google Scholar]
  31. Robitalle C, & Collin J (2016). Prescription psychostimulant use among young adults: A narrative review of qualitative studies. Substance Use & Misuse, 51(3), 357–369. 10.3109/10826084.2015.1110170 [DOI] [PubMed] [Google Scholar]
  32. Russell MA, Almeida DM, & Maggs JL (2017). Stressor-related drinking and future alcohol problems among university students. Psychology of Addictive Behaviors, 31(6), 676–687. 10.1037/adb0000303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Sattler S (2019). Nonmedical use of prescription drugs for cognitive enhancement as response to chronic stress especially when social support is lacking. Stress & Health, 35, 127–137. https://doi.org/smi.2846 [DOI] [PubMed] [Google Scholar]
  34. Schepis TS, Buckner JD, Klare DL, Wade LR, & Benedetto N (2021). Predicting college student prescription stimulant misuse: An analysis from ecological momentary assessment. Experimental and Clinical Psychopharmacology, 29(6), 580–586. 10.1037/pha0000386 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Schepis TS, Teter CJ, & McCabe SE (2018). Prescription drug use, misuse and related substance use disorder symptoms vary by educational status and attainment in U.S. adolescents and young adults. Drug and Alcohol Dependence, 189, 172–177. 10.1016/j.drugalcdep.2018.08.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Scott CK, Dennis ML, & Gustafson DH (2018). Using ecological momentary assessments to predict relapse after adult substance use treatment. Addictive Behaviors, 82, 72–78. 10.1016/j.addbeh.2018.02.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Shiffman S (2009). Ecological momentary assessment (EMA) in studies of substance use. Psychological Assessment, 21(4), 486–497. 10.1037/a0017074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stawski RS, Sliwinski MJ, Almeida DM, & Smyth JM (2008). Reported exposure and emotional reactivity to daily stressors: The roles of adult age and global perceived stress. Psychology and Aging, 23(1), 52–61. 10.1037/0882-7974.23.1.52 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sumstine S, Cruz S, Schroeder C, Takeda S, & Bavarian N (2018). Racial/ethnic variation in mental health correlates of substance use among college students. Journal of Ethnicity in Substance Abuse, 17(2), 94–107. 10.1080/15332640.2017.1300554 [DOI] [PubMed] [Google Scholar]
  40. Tam CC, Benotsch EG, & Weinstein TL (2020). Resilience and psychiatric symptoms as mediators between perceived stress and non-medical use of prescription drugs among college students. The American Journal of Drug and Alcohol Abuse, 46(1), 120–130. 10.1080/00952990.2019.1653315 [DOI] [PubMed] [Google Scholar]
  41. Tavolacci MP, Ladner J, Grigioni S, Richard L, Villet H, & Dechelotte P (2013). Prevalence and association of perceived stress, substance use and behavioral addictions: A cross-sectional study among university students in France, 2009–2011. BMC Public Health, 13, 724. 10.1186/1471-2458-13-724 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Tessner KD, Mittal V, & Walker EF (2011). Longitudinal study of stressful life events and daily stressors among adolescents at high risk for psychotic disorders. Schizophrenia Bulletin, 37(2), 432–441. 10.1093.schbul/sbp087 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Thrul J, Bühler A, & Ferguson SG (2014). Situational and mood factors associated with smoking in young adult light and heavy smokers. Drug and Alcohol Review, 33, 420–427. 10.1111/dar.12164 [DOI] [PubMed] [Google Scholar]
  44. van der Steen Y, et al. , Gimpel-Drees J, Lataster T, Viechtbauer W, Simons CJP, Lardinois M, Michel TM, Janssen MB, Bechdolf A, Wagner M, & Myin-Germeys I (2017). Clinical high risk for psychosis: The association between momentary stress, affective and psychotic symptoms. Acta Psychiatrica Scandinavica, 136(1), 63–73. 10.1111/acps.12714 [DOI] [PubMed] [Google Scholar]
  45. Wang C-CDC, & Castañeda-Sound C (2008). The role of generational status, self-esteem, academic self-efficacy, and perceived social support in college students’ psychological well-being. Journal of College Counseling, 11(2), 101–118. 10.1002/j.2161-1882.2008.tb00028.x [DOI] [Google Scholar]

RESOURCES