Moral Injury among Women Military Veterans and Demand for Cigarettes: A Behavioral Economic Investigation Using a Hypothetical Purchase Task

Abstract

Objectives:

Unlike the United States general population, veteran women – as opposed to veteran men – have greater smoking prevalence; yet, little is known regarding factors that influence smoking in veteran women. The purpose of this study was to begin examining the relationship between a psychological concept known as moral injury and demand for cigarettes among veteran women.

Methods:

Veteran women who smoke (n = 44) were recruited for this cross-sectional study from Amazon MTurk, Reddit, and a veteran-serving non-profit organization in June-July 2023. Consenting participants received $2 for completing the cigarette purchase task (CPT), Fagerstrom Test for Nicotine Dependence (FTND), and the military version of the Moral Injury Symptom Scale (MISS-M-SF). We examined five CPT demand indices and calculated a modified exponential demand model stratified by moral injury severity status (i.e., probable vs. unlikely).

Results:

Probable morally injured women exhibited significantly higher relative reinforcing value (RRV) for smoking than unlikely morally injured women (F_{1, 920} = 9.16, p = 0.003). Average cigarette consumption at $0 (i.e., $Q_0$) was 48.56% higher (M = 22.24 vs. M = 13.55) in probable compared to unlikely morally injured women (p = 0.04, Hedge’s g = 0.74). FTND scores were significantly correlated with ${\text{P}}_{\text{max}}$ (i.e., demand elasticity point) and ${\text{O}}_{\text{max}}$ (i.e., maximum expenditure) values in both populations (rs = 0.42 – 0.68, ps < 0.05).

Conclusions:

We provide preliminary evidence of the relatively high RRV of smoking in morally injured veteran women. Continued research is needed to refine the characterization of this relationship.

Approximately 21.6% veterans smoke cigarettes (Odani et al., 2018) compared to 12.5% in the general population (Cornelius et al., 2022). Many factors encourage military smoking, including targeted advertising and on-base price discounts (Haddock et al., 2009). Veterans also experience unique risk factors for smoking, including the psychological consequences of war (McDaniel et al., 2023). Therefore, characterizing nicotine dependence among veterans is important.

Women veterans are typically underrepresented in research even though they comprise 10% of the veteran population (Veterans Affairs, 2023). The VA is committed to reducing health disparities among women, as per the Women’s Health Research Network (Veterans Affairs, 2020). Given that women (28.9%) veterans have higher smoking prevalence rates than men (21.1%) veterans (Odani et al., 2018), we help meet VA goals by characterizing nicotine dependence and its correlates among women veterans.

A better understanding of smoking among women veterans may be advanced by using behavioral economic principles. Hypothetical purchase tasks (HPT) are used extensively in behavioral economics to understand addiction (Jacobs & Bickel, 1999). In HPTs, participants self-report the number of cigarettes they would smoke at various prices. As price increases, demand decreases and ultimately ceases at higher costs. HPTs are reliable and sensitive to established indicators of substance dependence (Zvorsky et al., 2019). The data derived from HPTs correspond closely to labor-intensive laboratory measures (Wilson et al., 2016).

Though HPTs are used extensively (Zvorsky et al., 2019), we were aware of only a single study using purchase tasks among veterans (Pebley et al., 2022). We also investigate a risk factor in our analysis that is relevant to the military, moral injury, which is defined as perpetrating or witnessing an event that transgresses moral beliefs (Litz et al., 2009). Moral injury can be distinguished from Post-Traumatic Stress Disorder (PTSD) because of its focus on morality and has been estimated to be present in 37% of veterans (Nichter et al., 2021). Moral injury is associated with many adverse behavioral health outcomes, including substance use disorders and suicide (McDaniel et al., 2023). Indeed, Maguen et al. (2023) showed that perpetration of a morally injurious act was associated with an 18% increase in the odds of having a substance use disorder (SUD). Despite this association of veterans with moral injury having greater odds of a SUD, to our knowledge, associations with nicotine dependence have not been explored. Therefore, the purpose of this study was to examine relationships between moral injury and nicotine dependence severity among women veterans.

Methods

Study Design and Sample

Participants were recruited from three sources in June and July of 2023: Amazon MTurk, Reddit, and a veteran-serving non-profit organization. Forty-six participants met the following inclusion criteria and received $2 for participation: (a) must identify as a woman, (b) must report having served in the U.S. Armed Forces, as confirmed by a DD214 date, and report their number of years served, and (c) must currently use cigarettes, as per an affirmative response to this question: “do you currently smoke cigarettes (yes/no)?” Participants were excluded for never having served in the U.S. Armed Forces, identifying as a man, and for reporting no current use of cigarettes. IRB approval was obtained from Southern Illinois University and all participants gave informed consent.

Measures

Several baseline sociodemographic characteristics were assessed, including age (years), race/ethnicity (White, Black/African American, or other race), annual income (≤$30,000, $30,001 to $59,999, and ≥$60,000), and educational attainment (≤high school diploma, some college/associate’s degree, and ≥4-year degree).

We used the FTND (Heatherton et al., 1991) to assess nicotine dependence. A screening question with a dichotomous response option (i.e., “do you currently smoke cigarettes (yes/no”) preceded the 6-item FTND. The FTND is widely used and has well-established reliability and validity (e.g., Pomerleau et al., 1994).

We used the Cigarette Purchase Task (CPT) to determine demand for cigarettes (Jacobs & Bickel, 1999). Instructions for the CPT prompt participants to assume that the cigarettes are their usual brand, they have the same income/savings that they have currently, they must consume the purchased cigarettes within 24 hours, and they can smoke without any restrictions. Participants were instructed to respond to the following questions: “how many cigarettes would you smoke if they were ____ each?” The blank in the previous question was filled in with escalating prices: $0, $0.02, $0.05, $0.10, $0.20, $0.30, $0.40, $0.50, $0.60, $0.70, $0.80, $0.90, $1.00, $2.00, $3.00, $4.00, $5.00, $10.00, $20.00, $80.00, and $160.00. We also provided the cost of a pack at each price. Then we calculated five demand indices, including intensity of demand (${\text{Q}}_0$), representing the number of cigarettes purchased at $0 or minimal cost, maximum expenditure the participant is willing to pay in 24-hrs (${\text{O}}_{\text{max}}$), the price at which ${\text{O}}_{\text{max}}$ occurs (${\text{P}}_{\text{max}}$), overall sensitivity to price increases ($\mathrm{\alpha }$), and breakpoint, representing the price at which consumption reaches zero.

Participants also completed the Moral Injury Symptom Scale Military Version Short Form (MISS-M-SF; Koenig et al., 2018), a 10-item instrument that measures psychological and spiritual symptoms of moral injury (Chesnut et al., 2022, p. 3389). The response options include a Likert-scale ranging from (1) strongly disagree to (10) strongly agree, except for the final question about religious faith, which includes options ranging from (1) strengthened a lot to (10) weakened a lot. The scale has acceptable reliability ($\mathrm{\alpha }=0.87$; Koenig et al., 2018). Total scores on the MISS-M-SF range from 10–100. As in previous studies with 100 point moral injury scales, we used a cut-off score of ≥36 to distinguish between not-clinically significant and clinically significant (i.e., functional impairment requiring treatment) symptomology (Mantri et al., 2020).

Statistical Methods

We applied Stein et al.’s (2015) algorithm to identify nonsystematic response types, which identified two participants who met criteria - reducing the total number of participants with useable data from 46 to 44. To calculate the empirical demand curve, we employed Koffarnus et al.’s (2015) modified exponential demand model. Demand curves were estimated for participants with unlikely and probable moral injury separately, as well as in the aggregate. Koffarnus et al.’s (2015) demand model is presented in equation 1:

$$Q=Q_0\ast {10}^{k\left(e^{-\propto Q_{0}C}-1\right)}$$ (1)

where $Q$ is consumption of cigarettes at price $C$, $Q_0$ is a person’s consumption when the price is $0, k represents the span of the function (i.e., 2), and $\mathrm{\alpha }$ represents demand elasticity. As in Cassidy et al. (2020), we calculated an extra sum-of-squares $F$-test to compare the CPT demand curves for participants with unlikely versus probable moral injury. After computing $Q_0$, $\mathrm{\alpha }$, ${\text{P}}_{\text{max}}$, ${\text{O}}_{\text{max}}$, and breakpoint for the two-subpopulations, we examined whether differences in these dependent variables held after adjustment for length of military service. Also, we determined convergent validity by calculating Pearson correlation coefficients between FTND scores and the five behavioral economic demand indices.

Results

Participant Characteristics

Thirty two of the 44 participants [72.73% (95% CI = 59.57 – 85.89)] met criteria for probable moral injury. Across moral injury status categories [i.e., unlikely moral injury (UMI) or probable moral injury (PMI)], the sample was primarily White (UMI = 66.67% vs. PMI = 59.38%; p = 0.44; entire sample = 61.40%), reported annual incomes of $30,000 to $59,999 (UMI = 50.00% vs. PMI = 53.13%; p = 0.79; entire sample = 52.30%), and reported at least a bachelor’s degree (UMI = 75.00% vs. PMI = 56.25%; p = 0.39; entire sample = 61.40%). Those with PMI had an average age of 42.59 years (SD = 12.73), while those with UMI had an average age of 37.50 years (SD = 37.50); however, these differences were not significant (t = 1.25, p = 0.22; M_{(entire sample)} = 41.20). The average length of service differed (t = 3.00, p = 0.005) in the two sub-populations by 9.24 years (UMI = 7.17 years vs. PMI = 16.41 years). Given that length of service differed in the two sub-groups, we adjusted the demand indices analyses based on this characteristic.

CPT Demand

Cigarette demand decreased as an orderly function of increasing price (Figure 1), with the Koffarnus et al. (2015) demand equation providing satisfactory fit to the data (R² = 0.85). Coefficients of determination were also acceptable in the UMI (R² = 0.86) and PMI (R² = 0.78) moral injury sub-groups, with demand in the latter shifted upward compared to the former (F_{1, 920} = 9.16, p = 0.003). Some participants – particularly in the PMI group – continued to hypothetically make cigarette purchases even at some of the highest prices, reflecting severe nicotine addiction (Figure 1).

Figure 1:

Hypothetical Demand Curves for Cigarettes among United States Women Veteran Smokers in the (A) Aggregate and (B) by Moral Injury Severity in 2023

Results of analyses – adjusted for length of service – on five demand indices showed that $Q_0$ and ${\text{O}}_{\text{max}}$ differed by moral injury status (Table 1). At $0, those with PMI purchased 22.24+2.41 cigarettes while those with UMI purchased 13.55+2.44 cigarettes (p = 0.04). Additionally, those with PMI spent significantly more money on cigarettes (${\text{O}}_{\text{max}}$, p = 0.03). Average ${\text{P}}_{\text{max}}$, breakpoint, and alpha, were not significantly different between moral injury categories after adjustment for length of service.

Table 1:

Behavioral Economic Demand Indices from the Cigarette Purchase Task among United States Women Veterans by Moral Injury Severity in 2023

Index	Unlikely Moral Injury (n = 12)		Probable Moral Injury (n = 32)		P	Hedge’s g
	M	SD	M	SD
${\text{Q}}_0$	13.55	8.11	22.24	12.76	0.044	0.74
$\text{α}$	0.01	0.06	0.02	0.09	0.968	−0.12
${\text{O}}_{\text{max}}$	6.29	4.14	17.29	13.58	0.031	0.93
${\text{P}}_{\text{max}}$	1.20	0.41	2.76	2.28	0.073	0.79
Breakpoint	13.69	27.58	29.13	35.13	0.443	0.46

Note. All p-values are adjusted for length of military service.

FTND scores were significantly associated with ${\text{O}}_{\text{max}}$ (r = 0.68, p = 0.04) and ${\text{P}}_{\text{max}}$ (r = 0.68, p = 0.04) among those with UMI, as well as among those with PMI [${\text{O}}_{\text{max}}$ (r = 0.42, p = 0.03); ${\text{P}}_{\text{max}}$ (r = 0.45, p = 0.03)], supporting convergent validity of the CPT with women veterans. Pearson correlation coefficients for other demand indices and FTND scores were not significant.

Discussion

Results from this study in women veterans showed that demand curves fit the data and described the observed sensitivity to price increases well. Veterans with PMI demonstrated greater levels of demand on two of five CPT indices. Hall et al.’s (2021) systematic review of moral injury research identified correlations between moral injury and PTSD, depression, anxiety, suicide, and SUD. Most of these studies reported positive correlations with moral injury. The present study adds to the literature by identifying a significant association between moral injury and nicotine dependence severity.

In a meta-analysis of HPT indices, ${\text{Q}}_0$ and ${\text{O}}_{\text{max}}$ were identified as the most sensitive measures of demand (Zvorsky et al., 2019). In >80% of the studies reviewed, ${\text{Q}}_0$ and ${\text{O}}_{\text{max}}$ were correlated with other SUD measures. Consistent with these findings, ${\text{Q}}_0$ and ${\text{O}}_{\text{max}}$ in the present study were sensitive to differences in moral injury severity. Furthermore, Higgins et al. (2022) reported that participants with greater cumulative vulnerabilities for chronic smoking, including affective disorders, also had greater demand for cigarettes, as measured by ${\text{Q}}_0$. The present study expands the list of risk factors that are relevant to cigarette demand. The significant associations between FTND severity and ${\text{O}}_{\text{max}}$ and ${\text{P}}_{\text{max}}$ further add to the growing literature supporting the validity of the CPT (Higgins et al., 2017; Zvorsky et al., 2019).

A few limitations of the present study merit mention. The data were cross sectional and thus didn’t support causal inferences. Although our study sample was small, this is not uncommon in studies using the CPT (MacKillop et al., 2008). Furthermore, trauma exposures during military service were not assessed, which could potentially influence differences in moral injury. Additionally, we did not include measures concerning age of first cigarette use, nor social (e.g., friends, family, co-workers) influences on smoking. Future studies should seek to better understand how these issues effect the relationship between moral injury and cigarette use. Lastly, we were unable to make gender comparisons on the CPT, as inclusion criteria for this study required participants to identify as a woman. Although the literature suggests that women veterans smoke at higher rates than men, clarifying the moderating role of gender on moral injury and smoking among veterans would be novel. Despite these limitations, this study provides important initial support for the association between moral injury severity and cigarette demand.

To our knowledge, this is the first study to provide evidence of the relatively high RRV of smoking in morally injured women veterans, indicating a potentially important association between moral injury, cigarette demand, and nicotine dependence severity. Continued research on the relationship between moral injury and cigarette use among women veterans is needed in order to better characterize the findings of this study.

Highlights.

• Recent data suggests that women veterans smoke at greater rates than men

• Moral injury is common in veterans, but has not been linked to smoking yet

• We examined demand for cigarettes among women veterans by moral injury status

• Women with probable moral injury had significantly higher demand for cigarettes

• Future research with larger samples is needed to characterize this relationship