Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Apr 1.
Published in final edited form as: J Behav Med. 2017 Oct 11;41(2):221–231. doi: 10.1007/s10865-017-9893-4

The longitudinal, bidirectional relationships between parent reports of child secondhand smoke exposure and child smoking trajectories

Ashley H Clawson 1, Elizabeth L McQuaid 2, Shira Dunsiger 3, Kiera Bartlett 4, Belinda Borrelli 5
PMCID: PMC5844796  NIHMSID: NIHMS912275  PMID: 29022139

Abstract

This study examines the longitudinal relationships between child smoking and secondhand smoke exposure (SHSe). Participants were 222 parent-child dyads. The parents smoked, had a child with (48%) or without asthma, and were enrolled in a smoking/health intervention. Parent-reported child SHSe was measured at baseline and 4, 6, and 12-month follow-ups; self-reported child smoking was assessed at these points and at 2-months. A parallel process growth model was used. Baseline child SHSe and smoking were correlated (r = 0.30). Changes in child SHSe and child smoking moved in tandem as evidenced by a correlation between the linear slopes of child smoking and SHSe (r = 0.32), and a correlation between the linear slope of child smoking and the quadratic slope of child SHSe (r = −0.44). Results may inform interventions with the potential to reduce child SHSe and smoking among children at increased risk due to their exposure to parental smoking.

Keywords: parent, child, smoking, secondhand smoke exposure, longitudinal

INTRODUCTION

Parental smoking, the most common source of child secondhand smoke exposure (SHSe) (Ding et al. 2010), poses immediate risk for the parent and compounded risk for youth due to the health effects of SHSe and the increased risk of smoking uptake (U.S. Department of Health and Human Services 2006). About 41% of children aged 3–11 and 34% of children aged 12–19 are exposed to SHS (Homa et al. 2015). Parental smoking plays an important role in youth smoking, and has been linked to adolescent intentions to smoke, smoking initiation, early onset, rapid escalation, and persistent smoking, with longer parental tobacco exposure related to increased risk (Chassin et al. 2008; Fuemmeler et al. 2013; Kandel et al. 2015; Leonardi-Bee et al. 2011; Mays et al. 2014; Peterson et al. 2006; Schuck, Otten, Kleinjan, et al. 2013; Vuolo and Staff 2013; Weden and Miles 2012). This risk appears to be modifiable; smoking initiation rates are lower among children whose parents quit smoking (den Exter Blokland et al. 2004; Otten et al. 2007; Vuolo and Staff 2013).

Studies that examine the relationships between parent and child smoking hypothesize that parental modeling of smoking is associated with offspring’s observational learning and increased likelihood to initiate smoking, consistent with social ecological theories (Bandura 1986, 2004). These studies utilize assessments of parents’ smoking history and/or current smoking status as indicators of parental modeling of smoking, rather than assessing children’s proximity to parental smoking. Though children are generally aware of parental smoking (Harakeh et al. 2006), the use of parent smoking status as a proxy for parental modeling fails to capture whether children witness the smoking behavior, therefore engaging in observational learning. Measuring children’s SHSe may be a satisfactory proxy of parental modeling of smoking because SHSe would reflect exposure to parental smoking. Though less prevalent than studies that analyze parent current smoking status as a risk factor for child smoking, some studies have examined the relationship between child SHSe and child smoking. A systematic review found that SHSe was associated with child smoking initiation and current smoking (Okoli and Kodet 2015). More research is needed to further understand the longitudinal associations between children’s SHSe and smoking patterns.

A child’s behavior also has the potential to influence other family members’ behavior (Bandura 1986, 2004); yet only one study has examined how child smoking predicts later parental smoking (Schuck, Otten, Engels, et al. 2013). Shuck et al. conducted a longitudinal study that examined the cross-lagged associations between self-reported child smoking and parental smoking: more smoking among children predicted more subsequent smoking among parents (Schuck, Otten, Engels, et al. 2013). The authors propose a family smoking contagion effect, i.e., more smoking among one family member is associated with increased smoking among other family members (Schuck, Otten, Engels, et al. 2013).

Most parents who smoke do not want their children to smoke (Bottorff et al. 2013; Tilson et al. 2005); therefore, some parents who smoke may respond to concerns about their child smoking by reducing the amount they smoke in the presence of their child (thereby reducing modeling and SHSe). For example, in a study that examined the initial effectiveness of an intervention to help parents who smoke restrict children’s access to parental tobacco, parents reported reductions in their child’s SHSe (Robinson et al. 2015). Because this was not a treatment target, it was hypothesized that the intervention may have increased parental concern about youth smoking, which in turn led to changes in parental smoking behavior. A recent trial found that providing parents who had called a Quitline and quit smoking for 24 hours with a program focused on preventing offspring smoking improved parental abstinence rates (Jackson et al. 2016). Thus, parents may react to concerns about offspring smoking by changing their own smoking behaviors in an effort to deter smoking among their children.

The current study uses data from a larger smoking cessation induction trial to examine the longitudinal, bidirectional relationships between child self-reported smoking and parent reported child SHSe. It was hypothesized that longitudinal changes in parent-reported child SHSe and child self-reported smoking would be associated, i.e., these processes would change in tandem. Across individuals, processes can change together over time in multiple patterns, e.g., one process increases while another decreases, both processes increase, etc. For the present study, two change patterns were hypothesized: 1) that increases in parent-reported child SHSe would be associated with increases in child self-reported smoking (i.e., a positive association between slopes), and 2) that increases in child self-reported smoking would be associated with decreases in parent-reported child SHSe (i.e., a negative association between slopes). The former hypothesis is based on the literature on parental smoking and modeling (Chassin et al. 2008; Fuemmeler et al. 2013; Kandel et al. 2015; Leonardi-Bee et al. 2011; Mays et al. 2014; Okoli and Kodet 2015; Peterson et al. 2006; Schuck, Otten, Kleinjan, et al. 2013; Vuolo and Staff 2013; Weden and Miles 2012) and smoking contagion (Schuck, Otten, Engels, et al. 2013); the latter hypothesis is based on data that parents who smoke may reduce their smoking around their child for fear of the child increasing smoking (Jackson et al. 2016; Robinson et al. 2015). Results from this study may enhance the understanding of the relationships between children’s passive and active smoking and inform tobacco interventions that have the potential to reduce both child SHSe and smoking.

METHODS

Participants

This is a secondary analysis of a subset of data from a smoking cessation induction/health education study for caregivers who smoke (BLINDED). The present study includes 222 caregiver-child dyads families with a child from 8–17 years old. Our sample includes children with asthma (hereafter referred to as the asthma group; n = 107) and without asthma (hereafter referred to as the healthy child group; n = 115).

Caregivers of children with asthma were recruited primarily from emergency departments and urgent care; caregivers of healthy children were recruited from community events and publicity. Caregivers were eligible for participation if they met these requirements: smoked ≥3 cigarettes per day for the last year and had smoked at least 100 cigarettes, were the primary caregiver of a child between the age 3–17, ≥18 years old, were not pregnant or planning to become pregnant, were reachable by telephone, were fluent in English, and were not enrolled in cessation treatment, using medication or nicotine replacement therapy to quit smoking. Caregivers of children with asthma were eligible for participation if their child had experienced an asthma exacerbation in the last two months necessitating urgent care (urgent care visit or hospitalization). In the asthma group, families with target children who had other significant respiratory illnesses were excluded; in the healthy child group, families with a child with asthma or significant respiratory illness were excluded. Participants did not have to want to quit smoking to enroll but had to be willing to discuss smoking and have home-based health education visits.

All participants completed two home visits that involved health education and motivational interviewing for smoking. Participants in the asthma groups received health education focused on asthma; the healthy group received education on general child health. After the home visits, participants in the asthma group (n=341) were randomized to receive one of two types of six counseling calls over four months: guidelines-based asthma education (National Asthma Education and Prevention Program 2007) plus child wellness (n=171) or the same asthma education plus a motivational intervention for smoking (n=170). After the home visits, parents in the healthy child group received six counseling calls focused only on child wellness (n=219). Only families with children ≥8 years old reported on child smoking; thus the current sample is comprised of 222 families with a child from 8–17 years old (asthma groups- n = 107; healthy child group- n = 115). The present sample enabled us to examine the study aim among a heterogeneous population, including youth at greater risk for tobacco-related consequences; however, power limitations precluded our ability to compare outcomes based on asthma status. Participants who wanted to quit within 30 days were provided with an 8-week supply of Transdermal Nicotine Patch treatment at no cost. Additional details of the larger study can be found in (BLINDED). The study was approved by our Institutional Review Board. Data were collected in Rhode Island and Massachusetts from 2008–2013. Informed consent was obtained from all individual participants included in the study. More detailed information about the study design can be found in (BLINDED REF).

Measures

Demographic (age, race, gender, education, income) and caregiver smoking history variables (cigarettes smoked per day, Fagerstrom Test for Nicotine Dependence (Heatherton et al. 1991), home smoking ban (no smoking allowed in the home), presence of other smokers in the home) were assessed via caregiver self-report at baseline.

Parent Reports of Child Secondhand Smoke Exposure (SHSe)

Assessments for parent-reported child SHSe were completed at baseline and 4, 6, and 12-months. Parent-reported child SHSe was assessed with a structured caregiver interview that produced a composite score of the number of cigarettes a child was exposed to from all people (excluding exposure from a child’s personal smoking) and all places during the past week. Higher scores indicated higher SHSe. The composite score has demonstrated good reliability and validity (Matt et al. 2000). Objective SHSe was measured at baseline with a passive nicotine air monitor (i.e., dosimeter) that the child wore for one week. Dosimeters have good validity (Leaderer and Hammond 1991). They were only used in the present study to establish the validity of parent-reported SHSe. Baseline assessments of objective child SHSe and parent-reported SHSe were correlated, r = .28, p < .001. Parent-reported child SHSe was used as a dependent variable because it excludes children’s SHSe from personal active smoking; SHSe as measured by passive dosimetery could not delineate parent vs. child smoking. Additionally, there were two more assessment points for parent-reported child SHSe, allowing for an examination of non-linear change. Participants received $20 per completed questionnaire and $10 for returning the dosimeters in good condition.

Child Smoking Status

Children were queried about their smoking status by an interviewer without the parent present at baseline and at 2, 4, 6, and 12-months. Smoking status was determined via validated questions about smoking quantity and frequency (Udry 2003). A smoking status composite score was compiled with higher scores indicating greater smoking and/or smoking-related exposure: (1) never smoked, (2) tried or puffed a cigarette, (3) had smoked cigarette, but had not smoked a 100 cigarettes and had not smoked during the past month, (4) had smoked a cigarette and had smoked 100 cigarettes, but had not smoked during the past month, (5) had smoked a cigarette during the past month, but had not smoked 100 cigarettes, and (6) had smoked 100 cigarettes and smoked during the past month.

Bioverified Parent Quit Status (7 day point prevalence abstinence (PPA))

At follow-up visits, self-reported caregiver quit status, (i.e., self-report of no smoking in the past 7 days), was verified using carbon monoxide testing (Bedfont, CO Ecolyzer). Self-reported quit was confirmed if readings were ≤ 9 ppm and readings > 9ppm were recoded as smokers (Benowitz et al. 2002).

Data Analysis

Analyses were performed with Mplus. Outliers were identified via graphical techniques; five cases with outlying values for SHSe were removed. SHSe was non-normally distributed with large variance so a square-root transformation was used. Standardized results are presented. We used latent growth curve modeling to examine how longitudinal changes in child smoking and child SHSe were related. This type of analysis is a recommended approach for studying youth smoking (Darling and Cumsille 2003). First, separate latent growth curve models were fit to child SHSe and child smoking to establish the best fitting growth curve for each process. Second, a parallel process growth model was used to examine the correlations between the latent growth factors (intercepts and slopes) of parent-reported child SHSe and child self-reported smoking. The final model included baseline child age and time-varying bioverified parent quit status (7 day PPA). Study group was included in an earlier model but due to significant correlation with child age, it was removed from the final model in order to avoid multicollinearity. Model fit was estimated with the χ2, Comparative Fit Index (CFI), Standardized Root Mean Square Residual (SRMR), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC) (Geiser 2013; Hu and Bentler 1999).

RESULTS

Participant characteristics are provided in Table 1. Correlations and descriptive statistics are provided in Table 2. First, separate latent growth curve models for parent-reported child SHSe and child self-reported smoking were tested; the best fitting growth curve was quadratic for child SHSe and linear for child smoking (Table 3). Next, a parallel process growth model concurrently examined the trajectories of parent-reported child SHSe and child self-reported smoking and the correlations between the latent growth factors of each process (Figure 1). Model fit was satisfactory (Table 3) (Geiser 2013; Hu and Bentler 1999). The final model controlled for baseline child age and time-varying parent bioverified quit status.

Table 1.

Participant Characteristics at Baseline (n = 222).

M (SD) n (%)
Parent Age 39.95 (8.88)
Parent sex (Female) 193 (86.9%)
Parent race/ethnicity
 White, non-Hispanic 112 (50.5%)
 Black, non-Hispanic 58 (26.1%)
 Hispanic 30 (13.5%)
 Other 22 (9.9%)
Parent- completed education beyond high school 93 (41.9%)
Income below $25,000 139 (62.6%)
Parent cigarettes smoked per day 15.55 (9.62)
Parent- Fagerstrom Test for Nicotine Dependence 4.77 (2.36)
Other smokers lived in the home 98 (44.1%)
Total home smoking ban in place 91 (41.0%)
Child Age 11.56 (2.98)
Child Sex (Female) 114 (51.4%)
Child Baseline Smoking Status (Never smoker)1 122 (86.5%)
Open in a new tab
1

This is the valid percent considering missing data.

Table 2.

Correlation matrix and descriptive statistics.

1 2 3 4 5 6 7 8 9 10 11
1. Baseline child smoking - 0.80*** 0.93*** 0.78*** 0.68*** 0.32*** 0.28** 0.39*** 0.08 0.30*** −0.06
2. 2-month child smoking - 0.69*** 0.83*** 0.78*** 0.28** 0.30** 0.37*** 0.04 0.35*** 0.07
3. 4-month child smoking - 0.86*** 0.84*** 0.09 0.11 0.26* −0.06 0.40*** 0.11
4. 6-month child smoking - 0.90*** 0.26** 0.22* 0.39*** 0.10 0.40*** 0.15
5. 12-month child smoking - 0.29** 0.30** 0.29** 0.09 0.24* 0.07
6. Baseline SHSe - 0.54*** 0.38*** 0.46*** 0.04 0.15*
7. 4-month SHSe - 0.62*** 0.61*** .08 0.18*
8. 6-month SHSe - 0.53*** 0.16 0.16*
9. 12-month SHSe - 0.05 0.19**
10. Child age - 0.20**
11. Study Group -
Mean 1.30 1.42 1.32 1.27 1.45 4.30 2.68 2.45 2.09 11.56 -
Standard Deviation 0.90 1.09 0.97 0.88 1.17 3.74 3.14 2.09 3.00 2.98 -
Open in a new tab

Notes: n’s range from 58–222. These n’s are small due to listwise deletion. In the Mplus models, the full information maximum likelihood (FIML) procedure was used to address missing data.

*

p <.05

**

p <.01

***

p <.001.

Table 3.

Results from the Latent Growth Curve Models (LGCM) for Child SHSe and Smoking.

Models χ2 (df) CFI SRMR BIC AIC
Child SHSe LGCM
1. Intercept only 110.23 (26)*** .71 .08 4060.24 4019.40
2. Lineara 53.47 (22)*** .89 .05 4028.17 3973.73
3. Quadraticb 24.51 (16) 0.97 .03 4031.95 3957.10
Child Smoking LGCM
4. Intercept onlyc 108.33 (42)*** .81 .10 1191.89 1149.08
5. Lineard 49.03 (36) .96 .05 1148.04 1085.46
6. Quadraticc,d 51.15 (30)** .94 .05 1177.17 1094.84
Parallel Process Growth Model
7. Quadratic Child SHSe and Linear Child Smoking b, e 136.36 (65)*** .91 .05 5189.30 5012.36
8. Quadratic Child SHSe and Linear Child Smoking (only controlling for child age and parent quit) b, e, f 125.16 (57)*** .90 .06 5150.58 5007.67
Open in a new tab

Notes: All models controlled for study group, child age (time invariant covariates) and parent quit status (time varying covariate), unless otherwise stated. A non-significant χ2 and the following values were indicative of good model fit: > .95 for CFI, < .08 for SRMR 32,33. A BIC change of 10 or more was interpreted as strong evidence for improved model fit 38.

*

p <.05,

**

p <.01,

***

p <.001.

a

Variance of the linear slope factor was constrained to 0.

b

The residual variances for the observed child SHSe were constrained to be equal.

c

Residual variance for the observed 4 month child smoking was constrained to 0..

d

Residual variance for the observed 12 month child smoking was constrained to 0.

e

Residual variances for the observed 4 and 12 month child smoking were constrained to be equal. The residual covariances were constrained to be equal. Autoregressive correlations were constrained to be equal.

f

Due to multicollinearity between child age and study group, model only controlled for child age and parent quit status.

Figure 1.

Figure 1

Open in a new tab

Conceptual diagram of parallel process growth model.

It should be noted that model fit and resulting patterns remained similar when study group (variable reflecting if participant was in the healthy group, asthma group with no additional smoking counseling, or asthma group with additional smoking counseling), was added to the model; however, risk of collinearity was high, and thus it was removed from the final model (Table 3). The pattern of results in the final model were also similar to a multiple group model that was stratified by healthy vs. asthma status (and controlled for child age, time-varying parent bioverified quit status, and study group (in the asthma portion of the model)), the correlations between slope growth factors were not significant, likely due to reduced power.

Growth Patterns of Parent-reported Child SHSe and Child Self-reported Smoking

Parent-reported Child SHSe

Baseline levels of parent-reported SHSe were different from zero (M = 1.04, p <.001) and there was significant variability (σ = 1.00, p <.001). There were linear decreases in parent-reported SHSe over time (M = −0.58, p <.001) followed by increases in parent-reported SHSe (M = 0.44, p =.001), the latter representing the quadratic growth over time. There was significant variability in the linear (σ = 0.99, p <.001) and quadratic growth terms (σ = 0.98, p <.001). Higher parent-reported SHSe at baseline was associated with fewer linear decreases in parent-reported SHSe (r = −0.79, p <.001) and more quadratic increases (r = 0.71, p <.001). The relationship between the linear and quadratic parent-reported SHSe slopes, commonly assessed as part of model interpretation, was significant (r = −0.98, p <.001), suggesting that more linear SHSe decreases were associated with fewer quadratic increases.

Child Self-reported Smoking

Baseline child smoking levels were different from zero (M = 1.36, p <.001) and had significant variability (σ = 0.88, p <.001). Child smoking increased over time (M = 0.31, p < .001), with significant variability in the linear growth term (σ = 0.98, p <.001). Higher baseline child smoking levels were modestly associated with more increases in child smoking over time, although this trend did not reach significance (r = 0.46, p = 0.08).

Correlations between Growth Factors for Parent-reported Child SHSe and Child Self- reported Smoking

Next, the correlations between the intercepts and slopes of parent-reported child SHSe and child self-reported smoking were examined. Higher baseline parent-reported child SHSe was associated with higher baseline child self-reported smoking (r = 0.30, p = 0.001). Baseline levels of parent-reported child SHSe were not correlated with the growth of child smoking across time (r = −0.004, p = 0.97). Baseline child smoking levels were not associated with linear or quadratic growth in parent-reported SHSe (r = 0.02, p = 0.88; r = −0.15, p = 0.27, respectively).

To examine the bidirectional, longitudinal relationships between growth in parent-reported child SHSe and child self-reported smoking over time, the correlations between each of the processes’ slopes were examined. The linear slope of child smoking and the linear slope of parent-reported child SHSe were correlated (r = 0.32, p = 0.04). This correlation indicates that these processes change in tandem: More changes in parent-reported child SHSe were associated with more changes in child self-reported smoking.

The positive correlation between child SHSe and smoking linear slopes identified that these processes change together; further examination of individual children’s slope values allowed for an exploration of how child SHSe and smoking changed across time for each participant. Each participant was mapped to a growth pattern (e.g., a linear smoking slope term and linear SHSe slope term was linked to each participant). Data from the scatterplot of the two linear slopes were examined to ascertain how each child changed in each process (Figure 2). We subdivided the plot of the two linear slopes into 4 quadrants and calculated the number of participants who fell within each quadrant; the mean slope values of child smoking and parent-reported SHSe within each quadrant were also examined. Results suggest the following patterns: 1) parent-reported child SHSe increased and child smoking decreased (n = 6, 2.7%), 2) parent-reported child SHSe increased and child smoking increased (n = 66, 29.7%), 3) parent-reported child SHSe decreased and child smoking increased (n = 95, 42.8%), and 4) parent-reported child SHSe decreased and child smoking decreased (n = 55, 24.8%).

Figure 2.

Figure 2

Open in a new tab

Scatterplot of linear slope values for child smoking and child SHSe.

The linear slope of child self-reported smoking and the quadratic slope of parent-reported child SHSe were also associated (r = −0.44, p = 0.02), indicating that more increases in child smoking were associated with fewer changes in the quadratic growth of parent-reported child SHSe. Participants moved in tandem via the following patterns: 1) child smoking increased with negative quadratic growth in parent-reported child SHSe (n = 77, 34.7%), 2) child smoking increased with positive quadratic growth in parent-reported child SHSe (n = 84, 37.8%), 3) child smoking decreased with positive quadratic growth in parent-reported child SHSe (n = 55, 24.8%), and 4) child smoking decreased with negative quadratic growth in parent-reported child SHSe (n = 6, 2.7%).

DISCUSSION

Rates of child SHSe and youth smoking are higher among families with a parent who smokes than in the general population (Homa et al. 2015; Vuolo and Staff 2013). Parental smoking is associated with increased adolescent intentions to smoke and smoking (Chassin et al. 2008; Fuemmeler et al. 2013; Kandel et al. 2015; Leonardi-Bee et al. 2011; Mays et al. 2014; Peterson et al. 2006; Schuck, Otten, Kleinjan, et al. 2013; Vuolo and Staff 2013; Weden and Miles 2012). It has been hypothesized that this is through the mechanism of observational learning or possibly ‘smoking contagion,’ where behavior observed within the family is modeled by other family members (Bandura 2004; Schuck, Otten, Engels, et al. 2013). Child smoking behavior also has the potential to impact parental smoking behavior; parents may respond to increases in child smoking by reducing the amount they smoke in the presence of their child (and thereby reducing SHSe). Our study identified that the longitudinal patterns of parent-reported child SHSe and child self-reported smoking were associated. Further, different patterns of how parent-reported child SHSe and child self-reported smoking changed across time together were found, including that 1) increases in child SHSe were associated with increases in child smoking and 2) increases in child smoking were associated with decreases in child SHSe. This study found that changes in child SHSe and smoking are related, highlighting the potential family-level health benefits of interventions that focus on breaking the intergenerational transmission of tobacco exposure and uptake among families with a parent who smokes.

This is the first study to examine how longitudinal changes in parent-reported child SHSe and child self-reported smoking are related over time. Our study design allowed us to effectively examine this aim because (1) changes in parent-reported child SHSe were considered likely during the intervention and (2) our longitudinal data allowed for a an examination of trajectories, which has been deemed as a more appropriate modeling approach for examining youth smoking as compared to static evaluations of smoking (Darling and Cumsille 2003). The current study used parent-reported child SHSe; this is likely a more accurate measure of child proximity to parental smoking than report of parental smoking status. This measurement approach also allows SHSe from the child’s own smoking to be excluded. As hypothesized, higher levels of baseline parent-reported child SHSe were associated with higher levels of child smoking, and parent-reported child SHSe and smoking changed in tandem over time.

Two change patterns were hypothesized: 1) increases in parent-reported SHSe would be associated with increases in child smoking and 2) increases in child smoking would be associated with decreases in parent-reported child SHSe. Because the correlations between the linear and quadratic slopes of parent-reported child SHSe and the linear slope for child smoking were significant, indicating that the processes change in tandem, further examination of individual change patterns was warranted; both of the hypothesized change patterns were confirmed. The most prevalent change pattern was that as self-reported child smoking increased, parent-reported child SHSe decreased. When comparing the linear slopes of child smoking and parent-reported child SHSe, this pattern was identified in around 43% of youth; a similar pattern was seen when examining the correlation between the linear child smoking slope and the quadratic parent-reported SHSe slope (about 35% of children had increases in smoking and continued decreases in SHSe). If parents are aware of their child’s smoking, this could indicate agreement with previous research that suggested that when parents know their child is smoking, or become aware of child risk for smoking, they reduce the amount of time they smoke in the presence of their child (Robinson et al. 2015). Parents may hope that if they smoke around their child less, it will be modeled less and the child will reduce their smoking. It is unclear from the present research whether parents were aware of their child’s smoking. Previous research has indicated that parental and child reports of child smoking have fair to substantial agreement (Kappa .55–.67) (Harakeh et al. 2006); however, child perceptions of adults knowledge of their smoking may be lower (Ditre et al. 2008). To confirm our hypothesis, future research could verify parental reasons for reducing their child’s SHSe.

An alternative hypothesis for our finding may be that as children started smoking more they started to spend more time away from home. This hypothesis seems plausible, although our analysis controlled for child age and this behavior is likely less feasible for younger children. One implication of these data may be that interventions could incorporate feedback to parents who smoke about their children’s increased risk of smoking to motivate parents to reduce child SHSe and prevent child smoking initiation/escalation. To date, only one intervention has aimed to change both child smoking and SHSe among children with parents who smoke. The intervention, a self-directed program for the parent-child dyad that focused on increasing antismoking socialization and home smoking bans among families with a parent who smokes, reduced offspring smoking initiation, but did not increase home smoking bans (Jackson and Dickinson 2003, 2006). A later evaluation of this program showed that it successfully prevented relapse among parents who had called a Quitline and were at least 24-hours quit (Jackson et al. 2016). These studies seem to support our hypothesis that offspring smoking status/risk can serve as a motivator for parental behavior change, with benefits for the parent and child.

Alternatively, our data could also imply that for some families, reducing SHSe would not have a positive effect on child smoking. One potential explanation for this could be the idea discussed by Darling and Cumsille (2003) that parental smoking creates an environment that makes smoking more likely for a child (through a combination of modeling, perceived acceptability of behavior, environmental factors), but there still needs to be a ‘trigger event’ (Darling and Cumsille 2003). Children with parents who smoke may be more likely to both encounter a trigger event and react to this event by increasing smoking than children with non-smoking parents. In this case, exclusively addressing parental smoking may not be sufficient to reduce child smoking as the susceptibility to a trigger event may remain. For example, an intervention that solely focused on promoting parental smoking cessation (without focusing on child smoking) was not effective at reducing child smoking onset (Schuck et al. 2015). Overall, parental smoking cessation is always a desired outcome, but some parents may not be ready to quit and interventions that address both child SHSe and smoking may be the most advantageous to reduce child risk of current and future tobacco exposure.

Additionally, in some cases, increases in parent-reported child SHSe were accompanied by increases in child smoking. Interestingly, our results also suggest that decreases in parent-reported child SHSe were associated with decreases in child smoking, suggesting that the risk incurred from parental modeling may be modifiable. These patterns of change were found when both the linear and quadratic slopes of parent-reported child SHSe were compared with the linear slope of child smoking and are commensurate with the concept of observational learning, and also smoking contagion, of smoking behavior among families. This suggests that SHSe reduction interventions, including interventions that promote parent cessation and/or aim to reduce child SHSe, have the potential to have a direct positive effect on child health through reductions in child SHSe and smoking. As proposed earlier, interventions that capitalize on the fact that child SHSe and smoking appear to change together and explicitly target both child SHSe and smoking may hold the most potential for breaking the intergeneration transmission of smoking. However, it is important to note that the current study did not examine whether the relationship between child smoking and SHSe was moderated by other variables that are risk factors for child smoking (e.g., access to tobacco); therefore, more research is needed to inform what types of interventions would be most effective.

Despite the contribution of this paper, there are limitations. While we were able to examine how changes in parent-reported child SHSe and smoking were correlated and provide hypotheses about why these change patterns were observed, we lacked data to confirm what factors influenced the identified patterns. This study could not control for other influences on SHSe and child smoking; future studies should examine other variables that may affect child SHSe or smoking. Due to multicollinearity and lack of power, we were unable to evaluate potential differences in these patterns based on child asthma status. Notably, the patterns of results were similar across the models that included study group and the multiple group model that was stratified by asthma status. Future studies could explore potential differences in these relationships between children with and without asthma. The fact that all parents were enrolled in a smoking intervention may limit the generalizability of our findings; however, the study was advertised as a health education intervention, parents did not have to want to quit smoking to participate, and the study design allowed us to effectively examine our hypothesis about tandem changes in parent-reported child SHSe and smoking, something that may be have been difficult in an observational study. Our study also had some limitations to our measurement. First, because parents reported on child SHSe and their own smoking behavior, common method variance may have influenced our results. Second, although measuring child SHSe was hypothesized to be a better proxy of parental modeling than parent smoking status alone, there is still uncertainty surrounding if the child witnessed the parent smoking. Third, the correlation between parent-reported child SHSe and objectively measured child SHSe was low (r = .28). This correlation reflects the association between parent reports of child SHSe based on all possible sources of SHSe (i.e., exposure from all people in all places) with the objective measure of child SHSe based on a dosimeter children wore for one week at baseline. This correlation may be low for several reasons. It is possible that the objective estimate of child SHSe underestimated the child’s exposure to SHS compared to parent reports because the dosimeter was not worn consistently or was placed somewhere where it could not efficiently gather nicotine. Alternatively, the low correlation may reflect parents’ difficulty with accurately reporting on their child’s exposure to SHSe for situations when the parent is not present. While our study provides preliminary support for a relationship between changes in child SHSe and smoking, future research should confirm these results by examining these relationships with measurement approaches less at risk for bias.

Conclusion

This is the first study to examine the longitudinal, bidirectional relationships between parent-reported child SHSe and child self-reported smoking. Parent-reported child SHSe and child smoking changed in tandem across time. The results may inform tobacco interventions that have the potential to reduce both child SHSe and child smoking among children at increased risk due to their exposure to parental smoking. Further research could aim to confirm the identified change patterns and explore variables that predict which patterns of change are most likely for families.

Acknowledgments

Funding: This work was supported by the National Institutes of Health (5 R01 HL062165-09 (B. Borrelli, PI) and 5 T32 HL076134-09 (R. Wing, PI)).

The authors thank Kristoffer S. Berlin, Ph.D. for his guidance.

Footnotes

Conflict of Interest: The authors have no conflicts of interest.

The study was conducted at The Miriam Hospital, when Dr. Borrelli was employed there. This study was completed when Dr. Clawson was affiliated with the Centers for Behavioral and Preventive Medicine and the Bradley/Hasbro Children’s Research Center. Dr. Clawson is now employed at Oklahoma State University.

Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

References

  1. Bandura A. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall; 1986. [Google Scholar]
  2. Bandura A. Health promotion by social cognitive means. Health education & behavior. 2004;31(2):143–64. doi: 10.1177/1090198104263660. [DOI] [PubMed] [Google Scholar]
  3. Benowitz NL, PJ, Ahijevych K, Jarvis MJ, Hall S, LeHouezec J, et al. Biochemical verification of tobacco use and cessation. Nicotine & Tobacco Research. 2002;4(2):149–159. doi: 10.1080/14622200210123581. [DOI] [PubMed] [Google Scholar]
  4. Bottorff JL, Oliffe JL, Kelly MT, Johnson JL, Chan A. Reconciling parenting and smoking in the context of child development. Qualitative health research. 2013;23(8):1042–53. doi: 10.1177/1049732313494118. [DOI] [PubMed] [Google Scholar]
  5. Chassin L, Presson C, Seo DC, Sherman SJ, Macy J, Wirth RJ, Curran P. Multiple trajectories of cigarette smoking and the intergenerational transmission of smoking: a multigenerational, longitudinal study of a Midwestern community sample. Health psychology: official journal of the Division of Health Psychology, American Psychological Association. 2008;27(6):819–28. doi: 10.1037/0278-6133.27.6.819. [DOI] [PubMed] [Google Scholar]
  6. Darling N, Cumsille P. Theory, measurement, and methods in the study of family influences on adolescent smoking. Addiction (Abingdon, England) 2003;98(Suppl 1):21–36. doi: 10.1046/j.1360-0443.98.s1.3.x. [DOI] [PubMed] [Google Scholar]
  7. den Exter Blokland EAW, Engels RCME, Hale WW, Meeus W, Willemsen MC. Lifetime parental smoking history and cessation and early adolescent smoking behavior. Preventive medicine. 2004;38(3):359–68. doi: 10.1016/j.ypmed.2003.11.008. [DOI] [PubMed] [Google Scholar]
  8. Ding D, Wahlgren D, Liles S, Jones J, Hughes S, Hovell M. Secondhand smoke avoidance by preteens living with smokers: To leave or stay? Addictive Behaviors. 2010;35(11):989–994. doi: 10.1016/j.addbeh.2010.06.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ditre JW, Coraggio JT, Herzog Ta. Associations between parental smoking restrictions and adolescent smoking. Nicotine & tobacco research: official journal of the Society for Research on Nicotine and Tobacco. 2008;10(6):975–83. doi: 10.1080/14622200802087549. [DOI] [PubMed] [Google Scholar]
  10. Fuemmeler B, Lee CT, Ranby KW, Clark T, McClernon FJ, Yang C, Kollins SH. Individual- and community-level correlates of cigarette-smoking trajectories from age 13 to 32 in a U.S. population-based sample. Drug and alcohol dependence. 2013;132(1–2):301–8. doi: 10.1016/j.drugalcdep.2013.02.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Geiser C. Data Analysis with Mplus. New York: The Guilford Press; 2013. [Google Scholar]
  12. Harakeh Z, Engels RCME, de Vries H, Scholte RHJ. Correspondence between proxy and self-reports on smoking in a full family study. Drug and Alcohol Dependence. 2006;84(1):40–47. doi: 10.1016/j.drugalcdep.2005.11.026. [DOI] [PubMed] [Google Scholar]
  13. Heatherton TF, Kozlowski LT, Frecker RC, Fagerstrom KO. The Fagerstrom Test for Nicotine Dependence: a revision of the Fagerstrom Tolerance Questionnaire. Addiction. 1991;86(9):1119–1127. doi: 10.1111/j.1360-0443.1991.tb01879.x. [DOI] [PubMed] [Google Scholar]
  14. Homa D, Neff L, King B, Caraballo R, Bennell R, Babb S, et al. Vital Signs: Disparities in Nonsmokers’ Exposure to Secondhand Smoke—United States, 1999–2012. MMWR. 2015;64(4):103–108. [PMC free article] [PubMed] [Google Scholar]
  15. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal. 1999;6(1):1–55. doi: 10.1080/10705519909540118. [DOI] [Google Scholar]
  16. Jackson C, Dickinson D. Can parents who smoke socialise their children against trial. Tobacco control. 2003;12:52–59. doi: 10.1136/tc.12.1.52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jackson C, Dickinson D. Enabling Parents Who Smoke to Prevent Their Children From Initiating Smoking. Results from a 3-year Intervention Evaluation. Archives of Pediatrics & Adolescent Medicine. 2006;160:56–62. doi: 10.1001/archpedi.160.1.56. [DOI] [PubMed] [Google Scholar]
  18. Jackson C, Hayes KA, Dickinson DM. Engaging Parents Who Quit Smoking in Antismoking Socialization of Children: A Novel Approach to Relapse Prevention. Nicotine & Tobacco Research. 2016;18(5):926–933. doi: 10.1093/ntr/ntv214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kandel DB, Griesler PC, Hu MC. Intergenerational Patterns of Smoking and Nicotine Dependence Among US Adolescents. American Journal of Public Health. 2015;105(11):e63–e72. doi: 10.2105/AJPH.2015.302775. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Leaderer B, Hammond S. Evaluation of vapor-phase nicotine and respirable suspended particle mass as markers for environmental tobacco smoke. Environmental Science and Technology. 1991;25:770–777. doi: 10.1021/es00016a023. [DOI] [Google Scholar]
  21. Leonardi-Bee J, Jere ML, Britton J. Exposure to parental and sibling smoking and the risk of smoking uptake in childhood and adolescence: a systematic review and meta-analysis. Thorax. 2011;66(10):847–55. doi: 10.1136/thx.2010.153379. [DOI] [PubMed] [Google Scholar]
  22. Matt GE, Hovell MF, Zakarian JM, Bernert JT, Pirkle JL, Hammond SK. Measuring secondhand smoke exposure in babies: The reliability and validity of mother reports in a sample of low-income families. Health Psychology. 2000;19(3):232–241. doi: 10.1037//0278-6133.19.3.232. [DOI] [PubMed] [Google Scholar]
  23. Mays D, Gilman SE, Rende R, Luta G, Tercyak KP, Niaura RS. Parental Smoking Exposure and Adolescent Smoking Trajectories. Pediatrics. 2014;133(6):983–991. doi: 10.1542/peds.2013-3003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. National Asthma Education and Prevention Program. Guidelines for the diagnosis and management of asthma. National Asthma Education and Prevention Program Expert Panel Report 3. Bethesda, MD: 2007. NIH Publication Number 08-5846. [Google Scholar]
  25. Okoli CTC, Kodet J. A systematic review of secondhand tobacco smoke exposure and smoking behaviors: Smoking status, susceptibility, initiation, dependence, and cessation. Addictive Behaviors. 2015;47:22–32. doi: 10.1016/j.addbeh.2015.03.018. [DOI] [PubMed] [Google Scholar]
  26. Otten R, Engels RCME, Van De Ven MOM, Bricker JB. Parental smoking and adolescent smoking stages: The role of parents’ current and former smoking, and family structure. Journal of Behavioral Medicine. 2007;30(2):143–154. doi: 10.1007/s10865-006-9090-3. [DOI] [PubMed] [Google Scholar]
  27. Peterson AV, Leroux BG, Bricker J, Kealey Ka, Marek PM, Sarason IG, Andersen MR. Nine-year prediction of adolescent smoking by number of smoking parents. Addictive Behaviors. 2006;31(5):788–801. doi: 10.1016/j.addbeh.2005.06.003. [DOI] [PubMed] [Google Scholar]
  28. Robinson La, Clawson AH, Weinberg Ja, Salgado-Garcia FI, Ali JS. Physician Intervention for Improving Tobacco Control Among Parents Who Use Tobacco. Clinical pediatrics. 2015;54(11):1044–50. doi: 10.1177/0009922814567304. [DOI] [PubMed] [Google Scholar]
  29. Schuck K, Otten R, Engels RCME, Barker ED, Kleinjan M. Bidirectional Influences Between Parents and Children in Smoking Behavior: A Longitudinal Full-Family Model. Nicotine & Tobacco Research. 2013;15(1):44–51. doi: 10.1093/ntr/nts082. [DOI] [PubMed] [Google Scholar]
  30. Schuck K, Otten R, Kleinjan M, Bricker JB, Engels RC. School-based promotion of cessation support: reach of proactive mailings and acceptability of treatment in smoking parents recruited into cessation support through primary schools. BMC Public Health. 2013;13(1):381. doi: 10.1186/1471-2458-13-381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Schuck K, Otten R, Kleinjan M, Bricker JB, Engels RCME. Promoting smoking cessation among parents: Effects on smoking-related cognitions and smoking initiation in children. Addictive Behaviors. 2015;40:66–72. doi: 10.1016/j.addbeh.2014.09.002. [DOI] [PubMed] [Google Scholar]
  32. Tilson EC, McBride CM, Brouwer RN. Formative development of an intervention to stop family tobacco use: the Parents and Children Talking (PACT) intervention. Journal of health communication. 2005;10(6):491–508. doi: 10.1080/10810730500228615. [DOI] [PubMed] [Google Scholar]
  33. U.S. Department of Health and Human Services. The health consequences of involuntary exposure to tobacco smoke: a report of the Surgeon General. Atlanta, GA: US …. Atlanta, Georgia; 2006. [PubMed] [Google Scholar]
  34. Udry J. The National Longitudinal Study of Adolescent Health (Add Health), Waves I & II, 1994–1996; Wave III, 2001–2002. Chapel Hill, NC: 2003. [Google Scholar]
  35. Vuolo M, Staff J. Parent and child cigarette use: a longitudinal, multigenerational study. Pediatrics. 2013;132(3):e568–77. doi: 10.1542/peds.2013-0067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Weden MM, Miles JNV. Intergenerational relationships between the smoking patterns of a population-representative sample of US mothers and the smoking trajectories of their children. American Journal of Public Health. 2012;102(4):723–731. doi: 10.2105/AJPH.2011.300214. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES