Smoking Patterns of Household Members and Visitors in Homes With Children in the United States FREE

VARIOUS ORGANIZATIONS that promote child health have made recommendations to decrease childhood exposure to environmental tobacco smoke (ETS).^{1 - 3} The American Academy of Pediatrics¹ advises physicians to routinely assess a child's exposure to ETS and to inform smoking parents about the dangers of ETS. The American Medical Association² and World Health Organization³ advise encouraging parents to quit smoking.

Environmental tobacco smoke has been associated with increased rates of acute lower respiratory tract infections,^{4 - 7} wheezing and asthma,^{5 ,8 - 9} otitis media,^{10 - 12} sudden infant death syndrome,^{13 - 15} medical visits,^{5 ,16 - 18} hospitalizations,^{5 ,19} and school absences in children.⁸ Healthy People 2000²⁰ set a goal of reducing the percentage of children (aged ≤6 years) exposed to ETS to 20% or less. The more recent Healthy People 2010²¹ set a stricter goal of 10% or less.

Concerns over the effects of ETS on children and adults have prompted diverse actions. The United States has seen a boom in clean indoor air ordinances. During the past decade, many of the high-profile battles about ETS have been waged over public spaces such as restaurants,²² bars,²³ work sites,²⁴ and airplanes.²⁵ As of March 2002, 437 clean indoor air ordinances had been passed in cities across the country.²⁶ As of December 2001, 24 states had banned smoking in day care centers or limited it to areas with separate ventilation.²⁷ Fifteen states have such restrictions for government workplaces, and 3 states (California, Utah, and Vermont) have such policies for restaurants.²⁷ On the federal level, the Pro-Children Act of 1994²⁸ prohibited smoking in all publicly funded schools. An executive order in August 1997 banned smoking in federal buildings.²⁹

Although such laws have been instrumental in limiting ETS in public spaces, they do not have a direct effect on ETS exposure in other locations. The home, for example, is the major site of ETS exposure for children.^{5 ,30} Because of the nature of their lungs and their breathing rates, children have higher relative levels of cotinine (a nicotine metabolite used as a marker of nicotine exposure) than adults after breathing comparable levels of ETS.³¹ Research findings are unclear on whether children exposed to smoking at home are more likely to take up smoking themselves.^{32 - 34}

In light of the risks that ETS exposure poses to children, it is important to determine how many children are exposed to smoke at home. Previous national studies^{35 - 36} have examined whether household members smoke in homes with children, but not whether visitors (eg, friends, relatives, and nannies) smoke there.

We found no studies based on data from a single survey that have provided a detailed profile of all smoking in US households with children. Likewise, we did not find nationally representative studies of the indoor smoking practices in households where no residents smoke at home. In this study, we use data from a nationally representative household-based interview survey to determine the prevalence and patterns of smoking by residents and visitors in homes with children in the United States.

The data come from the 1994 National Health Interview Survey (NHIS) and the 1994 NHIS Year 2000 Objectives supplement. The National Center for Health Statistics, Hyattsville, Md, administers these surveys. The NHIS is an annual survey that covers demographics, health, health care utilization, and insurance. The Year 2000 Objectives supplement covers topics related to the US Department of Health and Human Services' Healthy People 2000²⁰ health objectives, including tobacco use.

Trained survey administrators from the US Bureau of the Census, Washington, DC, conducted the interview survey in the home. A multistage sample design was used to represent the civilian noninstitutionalized US population.

Half of the households in the 1994 core NHIS sample were randomly assigned to receive the Year 2000 Objectives supplement. The tobacco items were administered to 1 randomly selected adult per family. The household response rate from the core NHIS was 94.1% (45 435 households), and the response rate for households eligible for the Year 2000 Objectives supplement was 84.5%, for an overall year 2000 response rate of 79.5%. The total sample for the supplement consisted of 19 738 families in 19 374 households; 6990 of these families had children. We use family and home interchangeably in this article.

The principal outcome variable for this study was smoking in the home. Some questions referred to household members' smoking practices, some to visitors' smoking practices, and some to the respondent's own smoking. The interviewer began by telling the respondent that the "next questions are about smoking inside this home," and then asked, "Does anyone who lives here smoke cigarettes, cigars, or pipes anywhere inside this home?" If the answer was yes, the respondent was asked, "On the average, about how many days per week do people who live here smoke anywhere inside this home?" The answers ranged from 0 to 7 days. To assess smoking by people other than residents, the respondent was also asked, "On the average, about how many days per week are there visitors who smoke anywhere inside this home?" Throughout the article, we refer to the combined outcome variable of residents or visitors smoking at least 1 day per week as regular smoking in the home.

Respondents were also asked about their smoking habits. For the purposes of the study, current smokers were defined as those who responded affirmatively when asked, "Have you smoked more than 100 cigarettes in your life?" and then responded "Everyday" or "Some days" when asked, "Do you now smoke cigarettes?"

We selected independent variables that had been found to be significant correlates of one of the outcome variables in at least one prior study.^{36 - 41} These variables include age of the respondent, number of parents or other adults in the household, race/ethnicity of the respondent, highest educational level of responsible adult family members, geographic region, size of community, annual family income, and age of the youngest child. The NHIS defines children as people aged 0 to 17 years (ie, <18 years).

We report univariate and bivariate statistics and results of logistic regression analyses. The data set includes variables for individual children (because a family can have >1 child) and for families. Therefore, we used the child and the family as units of analysis. Specifically, the family is the unit of analysis for Table 1 and for the second and third rows of Table 2; the child is the unit of analysis for the first row of Table 2.

The National Center for Health Statistics provides household weights and person weights in the core data set so national population-based estimates can be made. We consulted the National Center for Health Statistics about adjustment of weights for use with the Year 2000 Objectives supplement. We used the core household weight for family-level analyses and the core person weights for child-level analyses. Because the supplement was administered to a random sample of half of the core NHIS households, we multiplied the weights by 2 and adjusted for the supplement's nonresponse rate. Standard errors were computed with standard methods that incorporate the complex sampling design.⁴²

Regular smoking (≥1 d/wk) by residents occurs in 32% of homes in which children live. When smoking by visitors is included as well, the percentage increases to 36%, representing more than 11 million households with more than 21 million children. Sixteen percent of nonsmoking respondents with children in the home report that other residents or visitors smoke there. In 6% of homes where neither the respondent nor other residents smoke, visitors regularly smoke.

There is wide variation in the demographic characteristics of homes with children (Table 1). Homes with at least 2 adults have less regular smoking than single-parent households (33% in homes with at least 2 adults vs 43% in father-only homes and 46% in mother-only homes, P<.001). The amount of regular smoking in the home increases with decreasing age of respondent, adult educational level, family income, and metropolitan statistical area size. The amount of regular smoking in the home is highest for African American (41%) and white (38%) respondents compared with Hispanic (25%) and other (25%) respondents (P<.001).

In regression analysis, most bivariate patterns remain. However, the differences in respondent age (P = .41) and metropolitan statistical area size (P = .52) are no longer statistically significant. Also, the racial/ethnic differences change substantially, with African American (odds ratio, 0.59; P<.001) and Hispanic/Latino (odds ratio, 0.34; P<.001) respondents less likely than white respondents to report regular smoking in the home. Compared with respondents from the South, those from the West are less likely to have regular smoking in the home (odds ratio, 0.63; P<.001).

The data presented so far cover smoking in homes with children, but we also examined child-level data: 30% of children live in homes with smoking by residents at least 1 day per week, and 35% live in homes with any smoking by residents or visitors. To compare our results with those of a 1988 survey,⁴⁰ we looked specifically at children 5 years or younger and found that 34% live in homes with regular smoking.

In most homes with smoking, smoking occurs on a daily basis (Table 2). In 92% of homes with children where residents smoke at least once a week (accounting for 28% of all children), the residents smoke at home every day. In 25% of homes where only visitors smoke, visitors smoke every day.

Most smokers who live with children do not prevent all smoking in their homes. Among respondents who are smokers and who live with children, 85% report that regular smoking occurs in their homes.

Regular smoking (by residents or visitors) is less common in homes without children than in homes with children (32% vs 36%, P<.001). However, controlling for the independent variables that appear in Table 1, except for parents or adults in the household and age of youngest child, which do not apply to homes without children, in a regression analysis that includes all homes, the presence of children is not significantly associated with smoking in the home (P = .22).

Thirty-five percent of children in the United States live in homes where household residents or visitors regularly smoke—representing 21 million children living in 11 million homes—and in more than 9 of 10 cases, if household residents smoke at home, they smoke there every day. Overall, 28% of children in the United States (>19 million children) live in homes in which a resident smokes every day of the week.

We also found that most smokers do not restrict smoking to outside the home. Among respondents who are smokers and who live with children, 85% report regular smoking in their homes by residents or visitors. Studies using state-level data have found that many people restrict smoking at home, although there appears to be wide variation in the prevalence of such restrictions. In a 1993 telephone survey of Massachusetts adolescents, 25% who lived with an adult smoker reported that there was no smoking in their homes, and 22% reported that adults restricted smoking to specific rooms.³² Twenty-three percent of adolescents who lived with smokers reported that visitors were not allowed to smoke in the home. A telephone survey in 1996 and 1997 found that 76% of California residents (including 43% of smokers) banned smoking in the home and an additional 10% had some smoking restrictions in the home.⁴³ In another 1996 California telephone survey, 38% of smokers reported living in homes with a complete smoking ban, 26% reported partial home smoking restrictions, and the rest had no restrictions.⁴⁴ A 1997 telephone survey found that 60% of Connecticut residents did not permit smoking in their homes.⁴⁵

We found that 16% of nonsmoking respondents with children at home report regular smoking there by other residents or visitors. This represented 3.7 million families in the United States. In 6% of homes in which residents do not smoke, visitors smoke at least once a week. Prior national studies have not collected information on smoking in homes without smokers, although some of the statewide studies have looked at this issue. In the Connecticut study,⁴⁵ 38% of nonsmokers reported ETS exposure in their homes, and the Massachusetts study³² found that 54% reported a smoking ban for visitors.

Our study builds on the findings of several earlier national reports. A 1988 survey reported that 42% of children in the United States 5 years and younger were living in households with a smoker,⁴⁰ whereas our study showed a lower percentage (34%) of children in the same age range living in homes with regular smoking. One reason our number is lower may be that the former study measures whether smokers live in the home, not whether they actually smoke there. Other factors contributing to the differences could be an overall decrease in smoking since 1988,⁴⁶ an increase in awareness of ETS effects,⁴⁷ or a combination of the two. Parents may be making a greater effort to prevent their children's exposure to ETS. Although we do not have US data, Canadians increasingly believe that parents who spend time at home with small children should not smoke (70% in 1996, compared with 51% in 1992).⁴⁷

Another study³⁵ reported that about 15 million children were exposed to ETS in their homes. Our data found that many more children (about 21 million) lived in homes with regular smoking. The other study was not household-specific, but rather combined 3 different surveys to make estimates; calculations were based on households with adult smokers and thus omitted smoking by teenagers and visitors.

Smoking by visitors was also omitted in the estimates of a third study.³⁶ It found in 1988 to 1991 that 41%, 44%, and 36% of children aged 2 months to 3 years, 4 to 11 years, and 12 to 16 years, respectively, were living in homes where at least 1 household member smoked. These percentages are higher than our estimate of 35% of children aged 0 to 17 years who live in homes with ETS (from residents or visitors). There may have been some decline in home smoking rates over time with increased publicity about the risks of ETS for children, or there may have been differences in methods. Another study⁴⁸ using the 1991 NHIS found that 37% of children lived in homes with smoking.

We found that the prevalence of smoking in homes with children varies by demographic characteristics. When controlling for other characteristics, smoking in the home is less likely in homes with younger children, homes with at least 2 adults, higher parental educational level, higher annual family income, and location in the West compared with the South. In addition, African American and Hispanic/Latino respondents are less likely to have smoking in the home than white respondents.

A study⁴³ of California families reported that, among smokers who stated that most of their friends also smoked, the odds of having a ban on smoking in the home were more than 6 times higher for homes with children than for homes without children. By contrast, in our national sample, we found that the presence of children does not predict whether regular smoking occurs in the home. Compared with most states, though, California has been more aggressive in regulating smoking and educating the public about risks of ETS.^{49 - 50}

A strength of the NHIS is that it covers a nationally representative sample with a high response rate. As with any survey covering stigmatized behaviors, there could be underreporting of smoking in general or in the home. Respondents may have been embarrassed about their smoking habits and refrained from reporting them to the interviewer. Objective markers of ETS exposure (eg, cotinine levels and ambient room smoke levels) were not collected. However, several studies^{36 ,51 - 53} have found that responses to questions about one's own smoking tend to be valid. The tobacco items were incorporated into an extensive interview covering many less-sensitive health topics, so interviewers had time to develop a rapport with respondents. Also, it might have been difficult to hide smoking behavior during in-home interviews, if smoking paraphernalia (eg, ashtrays) were present.

There may be other sources of ETS exposure for children (eg, cars and homes of friends and relatives),^{43 ,54} and a recent study⁵⁵ found that smoking sometimes occurs in homes in which respondents report that smoking is banned.

Our study may have undercounted some homes with daily smoking. There could have been daily smoking if a resident and visitor each smoked several days per week but smoked on different days. However, we did not know which days people smoked, so we were unable to combine smoking by residents and visitors who smoked at least 1 day but fewer than 7 days a week.

Factors such as the size of homes and quality of room ventilation can affect how much a child is actually exposed to ETS. However, most organizations recommend restricting smoking to outside the home, because limiting smoking to rooms without children and to well-ventilated rooms does not eliminate a child's ETS exposure.

Our data from the 1994 NHIS show that we were falling far short of achieving the Healthy People 2000 and Healthy People 2010 goals of reducing children's ETS exposure. Although rates may have declined since this study, they were high enough that it is unlikely that children are no longer being exposed to ETS in their homes. Proposed strategies to reduce children's ETS exposure include policy-based restrictions, legal sanctions, media and educational efforts, and clinical interventions.^{5 ,56} Legislation and litigation have been used in many communities to promote smoke-free public buildings and work sites, but they have been used in only limited ways to regulate smoking in the home—primarily in child-custody cases^{57 - 58} and multiresident dwellings.⁵⁹ One study⁶⁰ found that smoking restrictions in the home were more strongly associated with successful smoking cessation efforts than workplace smoking bans.

Education may have greater potential to reduce ETS in the private home. Organizations like the US Environmental Protection Agency, Washington, and the American Heart Association, Dallas, Tex, have sponsored public education campaigns.² Pediatricians and other clinicians who take care of children can provide health education to parents.^{5 ,61} One study⁶² showed that many clinicians, especially pulmonologists and allergists, report taking a history of passive smoke exposure. However, other studies report that clinicians often do not counsel parents about strategies to reduce passive smoking exposure⁶² or stop smoking altogether.^{63 - 64} We do not have clear evidence on the effect of clinical interventions on smoking in homes with children. One randomized controlled study⁶⁵ showed that a 1- to 2-minute intervention by pediatricians in the context of well-baby care can increase maternal smoking cessation and have a positive effect on relapse prevention. Another randomized controlled study⁶⁶ did not show an effect of a 10- to 15-minute intervention on maternal smoking cessation but did show a sustained effect on where smoking occurred and did improve knowledge of the health effects of ETS on children. However, recent reviews^{56 ,67} of the literature on physician-patient counseling found that onetime clinical smoking interventions are marginally efficacious at best, but that repeated minimal interventions by physicians and nurses can lead to higher rates of smoking cessation. There is evidence that clinicians' efforts to reduce ETS exposure can improve the health of children, such as by the prevention of asthma.⁶⁸ Education regarding smoking counseling techniques can also benefit clinicians, especially those in training, by giving them confidence in their counseling skills and encouragement to counsel parents.^{69 - 71}

Although smoking cessation would have the most far-reaching benefits, some parents may be more receptive to counseling to reduce ETS exposure for their children than to stop smoking altogether. Several studies^{56 ,61 ,72} have shown decreased exposure to ETS among children whose parents have received repeated clinician counseling about the effects of ETS and suggestions for reducing exposure. In homes with children, 16% of respondents who do not smoke report regular smoking in the home by others. Therefore, we found that physicians who want to reduce smoking in their patients' homes must ask not only whether the parent accompanying the child smokes but also whether other residents and visitors smoke. More than 3 million households with ETS exposure would be missed if the clinician prematurely ended the tobacco screening after being assured that the child's accompanying parent does not smoke. Some homes where no residents smoke have regular smoking by visitors, so clinicians may need to address ETS with all families, regardless of whether the parent who accompanies the child is a smoker and regardless of whether there is a smoker living in the home.

Corresponding author and reprints: Mark A. Schuster, MD, PhD, RAND, 1700 Main St, Santa Monica, CA 90407-2138 (e-mail: schuster@rand.org).

Accepted for publication June 13, 2002.

This study was funded in part by grant U48/CCU915773 from the Centers for Disease Control and Prevention, Atlanta, Ga, to the UCLA/RAND Center for Adolescent Health Promotion, Santa Monica.

All analyses, interpretations, and conclusions are the work of the authors and not of the National Center for Health Statistics, which is responsible only for the initial data. In addition, this work does not necessarily represent the opinions of the funding agency or of the institutions with which the authors are affiliated.

We thank Amy M. Bastian, MPH, who performed some of the programming; J. Neil Russell, PhD, for technical support in analyzing the data; Baback B. Gabbay, BA, Janina L. Morrison, AB, and Katherine D. Vestal, BA, for research assistance; and Alaida Rodriguez, BS, for administrative assistance.

Despite the health risks of ETS, children in the United States are more likely to be exposed to ETS in their homes than anywhere else. Prior national studies have examined smoking in households with children, but have not offered a comprehensive view of all sources of ETS in the home, including parents, other residents, and visitors (eg, friends, relatives, and nannies). In this study of a nationally representative sample of 6990 homes with children, the authors report that 35% of children in the United States live in homes where household residents or visitors regularly smoke, representing 21 million children living in 11 million homes. In addition, 16% of nonsmoking respondents with children report regular smoking in their homes by other residents or visitors, revealing an understudied population of children at risk for ETS exposure and the associated negative health outcomes.