Objectives: Secondhand smoke (SHS) exposure is an important and preventable cause of mortality and morbidity among children; hospitalization represents a sentinel event that may offer opportunities for intervention. The goal of this study was to determine the frequency and validity of SHS exposure screenings by emergency department (ED) providers, residents, and nurses.
Methods: A total of 140 inpatient pediatric families consented to a salivary cotinine measurement, in-person SHS exposure interview, and chart review to assess ED provider, pediatric resident, and nurse SHS exposure screenings and documentation validity.
Results: ED providers documented screening for SHS exposure 46% of the time, pediatric residents 42% of the time, and nurses 79% of the time. ED providers, pediatric residents, and nurses reported 18%, 38%, and 12% of patients exposed to SHS, respectively, whereas 46% of patients were identified as smoke-exposed according to cotinine level and/or parent report. Those with SHS exposure outside the home were least likely to be identified as exposed.
Conclusions: The majority of smoke-exposed children were not identified as exposed based on documentation of admission screenings. Future research is important to identify accurate and efficient methods of screening for and identifying SHS exposure among children admitted to the hospital.
Secondhand smoke (SHS) exposure is an important cause of mortality and morbidity among children and is related to many negative and costly health outcomes, such as lower respiratory tract infections,1 asthma,2–4 otitis media,5 inflammatory bowel disease,6 metabolic syndrome,7 leukemia,8 and sudden infant death syndrome.9 Even children with low levels of exposure had significantly decreased scores on tests of cognition10 and lower levels of antioxidant micronutrients11 compared with those children with no exposure. It is estimated that parental smoking results in annual direct medical expenditures of $4.6 billion and loss-of-life costs of $8.2 billion.12
Children exposed to SHS are significantly more likely to have health care costs associated with respiratory illness than those who are not exposed.13 Many programs have targeted SHS exposure screening and reduction in outpatient settings, yet hospitalization may represent a sentinel event for families in which opportunities for intervention may be present. Previous studies have shown parental receptivity to interventions in the inpatient setting,14 especially when their children are admitted with respiratory illness.15 Although recommendations have been made to regularly screen for SHS exposure in medical settings,16 little has been done to examine the frequency and validity of admission screenings for SHS exposure within the hospital setting. To best assist parents in reducing SHS exposure for their children, providers must have a method to accurately identify SHS exposure among children in an inpatient setting.
The specific aims of the present study were to (1) determine the frequency with which emergency department (ED) providers, pediatric residents, and nurses documented screening for SHS exposure, and (2) determine the validity of admission screenings for SHS exposure compared with a structured caregiver interview and objective cotinine levels. We hypothesized that (1) ED providers, pediatric residents, and nurses would vary in the frequency with which they documented in the chart screening for and identifying smoke exposure, (2) a detailed, structured caregiver interview would more accurately identify SHS exposure compared with admission screenings, and (3) the sensitivity of provider screenings for SHS exposure would be better for children admitted with respiratory illness than for those with other diagnoses.
Research methods were approved by the Research Subjects Review Board at the University of Rochester Medical Center in Rochester, New York.
The study took place at the Golisano Children’s Hospital (GCH) at the University of Rochester Medical Center in Rochester, New York, a “children’s hospital within a hospital,” and the tertiary care center for the surrounding counties of upstate New York. GCH is an academic institution, with pediatric, family practice, and ED residents and a dedicated pediatric ED.
Participants were recruited from all children admitted to GCH on general pediatrics or on a pediatric subspecialty service. Parents were invited to participate in the study if their children had been admitted in the previous 24 hours. Those with no guardian available at the time of admission, who were in foster care, whose anticipated stay was <24 hours, and whose families did not speak English were excluded. Recruitment took place from April 2008 to June 2009.
New admissions were reviewed weekday mornings by the study team, and eligible patients were identified. With permission from the patient’s treatment team, a trained interviewer approached the family for study enrollment. The interviewer obtained consent for a salivary cotinine measurement, in-person SHS exposure interview, and chart review to assess ED provider, pediatric resident, and nurse SHS exposure screenings. After initially very low response rates (16%), we began to offer a $5 coffee cart gift card for participation, which increased our response rate to 49%. Details on patients not approached or who refused are not available.
Before beginning the structured caregiver interview, a study team member collected a saliva sample from the patient using cotton-tipped Salivettes (Salimetrics, Inc., State College, PA). Salivary cotinine was used because it has correlated well with reports of exposure and serum cotinine,17–19 is noninvasive, can be collected quickly and easily in young children, and is stable for storage and transport under ambient conditions.20 Samples were batched and shipped to the Children’s National Medical Center in Washington, DC, for analysis. Cotinine was measured quantitatively using the enzyme-linked immunosorbent assay technique and is reported in nanograms per milliliter, with a limit of detection of 1.0 ng/mL.21 Therefore, children with a cotinine value ≥1 ng/mL were identified as having been exposed to SHS. We attempted to collect all saliva samples within 24 hours of admission, as cotinine has a half-life of 17 to 20 hours.22 However, 26% of patients had their saliva collection delayed past 24 hours from admission due to their parents’ request, and the average time of collection was 32 hours 47 minutes after admission. There was no difference in cotinine level for those whose saliva was collected within 24 hours compared with those whose saliva was collected after 24 hours; all subjects with cotinine measurements were therefore included in the analyses.
Structured Caregiver Interview for SHS Exposure:
Participating caregivers took part in a previously validated, structured, in-person interview.23 This interview included questions about the caregiver’s smoking habits and efforts to quit, as well as the child’s exposure to SHS by other household members or visitors (including household smoking bans), in cars, and in other settings (eg, day care, friends’ or relatives’ houses, public places. etc). Patients who were exposed to SHS in the last 7 days and/or were living with a caregiver who smoked were considered to be exposed to SHS. We also created a variable which specified that there was “SHS exposure inside” (ie., in the child’s house or apartment), “SHS exposure other,” or “no SHS exposure.”
Combined SHS Exposure Variable:
A child was considered smoke-exposed by our combined SHS exposure variable if they lived with a smoker based on parent report, had been exposed to SHS in the previous 7 days, or had a cotinine level >1 ng/mL. This strategy was designed to identify as many smoke-exposed children as possible.
Admission Screenings for SHS Exposure Via Chart Review:
ED providers, pediatric residents, and nurses were unaware of the aims of the study. ED providers have a check box for “tobacco” “alcohol,” and “drugs” in the social history section of their standard visit form, with options for none, current, and quit. Pediatric residents are taught to document tobacco smoke exposure as part of the social history. Nurses have a specific question on their admission note (“Does anyone smoke in the house or around the patient?”); this question is followed by a prompt to provide cessation counseling.
Chart reviews allowed the study team to determine the frequency and outcome (no SHS screening, SHS exposure, or no SHS exposure) of documented SHS exposure screenings. Written reports of smoking location were inconsistent; therefore, exposure was not defined by location on the basis of admission screenings. The admission diagnosis was noted. Those admitted with pneumonia, reactive airway, asthma, respiratory syncytial virus, a respiratory tract infection, respiratory croup, bronchiolitis, pneumothorax, or breathing problems were categorized as having respiratory illness. Those with diagnoses unrelated to the respiratory system were categorized as having no respiratory illness. This categorization allowed the study team to assess differences in SHS exposure screenings on the basis of diagnosis type. In addition, time of presentation to the ED and admission were recorded to determine the length of time in the hospital before saliva sampling.
Because cotinine levels were not normally distributed, Kruskal–Wallis and Mann-Whitney U tests were used to determine associations between cotinine levels, parent report of SHS exposure, and provider screening. Sensitivities and κ statistics were used to assess the accuracy of SHS exposure screening data found in patient charts. Wilson confidence intervals (CIs) were used for proportions.24 All analyses were conducted by using SPSS 18.0 (SPSS Inc, Chicago, IL).25
Of 432 families approached, 140 agreed to participate in the study (32%). The first 21% (n = 30) of enrolled participants were not offered an incentive to participate. The last 79% (n = 110) were offered a $5 gift card for a coffee cart. Those offered an incentive did not differ from those not offered an incentive with regard to demographic variables or the outcome measures selected for this study.
The mean age of patients was 3.9 years (range 0–16.9 years); 56% were male, 75% white, 10% African-American, 7% multiracial, 6% Hispanic, and 1% Asian. Overall, 32% of the patients were admitted with respiratory illnesses.
SHS Exposure Screening Prevalence
ED providers screened for SHS exposure 46% of the time, pediatric residents 42% of the time, and nurses 79% of the time. Patients were screened for SHS exposure by at least 1 person 93% of the time, meaning that 7% of inpatients were never screened for SHS exposure by an ED provider, pediatric resident, or nurse. Of those screened, ED providers, pediatric residents, and nurses documented that 18%, 38%, and 12% of patients were exposed to SHS, respectively. Of the 93% of patients who were assessed by anyone for SHS exposure, 26% were found to be exposed by at least 1 provider.
SHS Exposure by Cotinine Level and Caregiver Interview
Of the 140 participants, salivary collection was refused after consent for 22 (16%) and an insufficient volume was obtained for 37 (26%), leaving valid cotinine measurements for 81 patients (58%). Of these, 28% of patients had cotinine levels ≥1 ng/mL, indicating exposure to SHS. Using the structured caregiver interview, 139 patients (99%) were identified as exposed to SHS (45%) or not exposed to SHS (54%). One family did not know whether the child was exposed (1%). Of those exposed, 12% were exposed inside homes and 34% in other locations. Figure 1 shows the percentage of all patients who were screened for SHS exposure, as well as the percent identified as exposed to SHS by diagnosis using all 3 types of SHS exposure measurement. Based on our exposure measure combining parent report of any exposure in the past 7 days or a cotinine level >1 ng/mL, 46% of children were classified as smoke-exposed. There were no differences according to age, gender, or race/ethnicity (Table 1).
Differences in Cotinine by Levels of SHS Exposure
Cotinine levels corresponded with parent report; the median cotinine level for those identified as not exposed was 0.28 ng/mL (range 0.00–0.93 ng/mL); exposed outside the home, 1.03 ng/mL (range 0.00–15.67 ng/mL); and exposed inside the home, 1.72 ng/mL (range 0.14–25.00 ng/mL). Mann-Whitney U tests were conducted to evaluate pairwise differences among the 3 groups; compared with those not exposed, children exposed outside the home and those exposed inside the home had significantly higher cotinine levels (P < .001). Median cotinine levels for those exposed to SHS inside the home did not differ significantly from those exposed outside the home (P = .26).
Significant differences in cotinine levels were also found when looking at admission screenings (SHS exposure, no SHS exposure, or no screening) by ED providers (χ 22 = 6.86; P = .03) and nurses (χ 22 = 9.25; P = .01). Patients identified by ED providers as exposed to SHS had higher cotinine levels (median 3.56) compared with those not exposed to SHS (median 0.39; P = .02) and those not screened by ED providers for SHS (median 0.47; P = .01). Patients identified by nurses as exposed to SHS had higher cotinine levels (median 2.11) than those not exposed to SHS (median 0.47; P = .003) and those not assessed by nurses for SHS exposure (median 0.30; P = .02). There were no significant differences in cotinine levels by SHS exposure screening by pediatric residents (χ 22 = 4.32; P = .12).
There were no significant differences in cotinine level based on the length of time between admission and cotinine collection, regardless of whether SHS exposure was identified inside the home, outside the home, anywhere, or not at all. We therefore did not control for time of cotinine collection for subsequent analyses.
Admission Screenings Versus SHS Exposure Levels
When comparing documentation of admission screening with our combined exposure measure, the ED provider screening had a sensitivity of 41.2% (95% CI: 21.6%–64.0%), pediatric residents had a sensitivity of 66.7% (95% CI: 48.8%–80.8%), and nurses had a sensitivity of 26.7% (95% CI: 16.0%–41.0%). Table 2 displays sensitivities and κ values for admission screening documentation compared with each level of SHS exposure for all patients, patients admitted with respiratory illness, and patients admitted without respiratory illness. Pediatric residents and nurses most often correctly identified SHS exposure when the exposure occurred inside the home (89% [95% CI: 57% to 98% and 33%, 14% to 61%, respectively]), whereas ED providers were more accurate in identifying smoke exposure when it occurred outside the home (42% [95% CI: 19%–68%]). Respiratory illness admissions overall had improved sensitivity (71% [95% CI: 36%–92%] for ED providers; 82% [95% CI: 52%–95%] for residents; and 40% [95% CI: 17%–69%] for nurses) over those admitted for other reasons.
A percentage of children exposed to SHS was not identified by providers as exposed either because they were not screened or because documentation of the screening failed to identify smoke exposure. Eighty-five percent of children exposed to SHS were not identified by ED providers. Of these, 71% were not screened for smoke exposure (24% exposed inside, 76% outside the home) and 29% were misclassified as not exposed (30% exposed inside, 70% outside the home; P = <.05). Residents failed to identify 69% of smoke-exposed children. Of these, 77% were never screened by the residents (21% exposed inside, 79% outside the home) and 23% were identified as not exposed (10% exposed inside, 90% outside the home; P < .001). Nurses did not record exposure for 81% of children who were exposed to tobacco smoke. Of these, 31% were not screened for smoke exposure (27% exposed inside, 73% outside the home) whereas 70% were screened but not identified as smoke-exposed (24% exposed inside, 76% outside the home; P < .001). There were 7% of children who were never screened by any provider; of these, 0% were exposed to SHS inside and 30% outside the home.
Structured Caregiver Interview Versus Cotinine Level
Table 3 presents sensitivities and weighted κ values for the structured caregiver interview compared with cotinine levels. There were no children identified as smoke-exposed according to cotinine level and not by the structured caregiver interview. There were, however, 18 patients (16.67% exposed inside, 83.33% exposed outside the home) whose caregiver reported SHS exposure but who had a cotinine value <1 ng/mL; this finding did not vary significantly by length of time before saliva sampling. Thus, the structured caregiver interview identified children with SHS exposure more often than cotinine values.
There is no safe level of SHS exposure for children, yet 7% of children admitted to the hospital in this study were never screened for SHS exposure by any provider; 30% of these were exposed to tobacco smoke outside of the home. In addition, whereas 46% of patients were exposed to SHS on the basis of parent report and, in many cases, a biological marker, documentation of hospital providers’ admission screenings identified only 24% as exposed. This gap represents a significant missed opportunity to provide smoking cessation and smoke-exposure reduction counseling for families. There are a number of potential reasons for these discrepancies. There is no standardized way to assess SHS exposure in inpatient settings, and there is little training on assessing smoke-exposure in the inpatient setting, especially when the exposure is not inside the home. In addition, the setting of admission (often in the middle of a busy night, after a long stay in the ED) may not seem conducive to an in-depth discussion of parental smoking habits.
Traditionally, tobacco use has been documented in the social history of the patient note; in many institutions, SHS exposure is treated similarly. However, our current ED forms do not distinguish between use and exposure, and thus reporting actual exposure may be misinterpreted or misused. For instance, a provider may indicate “no tobacco” in that section, actually meaning that the infant is not, in fact, a smoker. In our study, ED providers failed to identify 85% of children exposed to SHS; 29% of these were misclassified as not exposed to smoke. Pediatric residents may have more time to obtain a detailed social history; however, they still failed to identify 69% of exposed children. Residents were less likely to misclassify exposure for children with respiratory illness and those who were exposed to SHS inside the home.
Nurses, who had the most direct prompt to assess SHS exposure, had the highest rate of screening but the lowest sensitivity and failed to identify most smoke-exposed children. It is possible that the question “Does anyone smoke in the home, or around the child” is ambiguous enough that parents will answer no, and because the screening has been completed, there is no reason to delve further. An additional factor may be that nurses have a higher smoking prevalence (13%) than other health care providers.26 Health care providers who smoke are less likely to provide smoking cessation intervention; whether this fact also may account for lower rates of screening in inpatient settings is not known.
It is likely that, even though most caregivers are asked about SHS exposure at some point while their child is hospitalized, the detailed questions which clearly demonstrate the nature and extent of SHS exposure are not being asked. If the “right” questions are not being asked, opportunities to discuss ways to reduce SHS exposure for children and their families will be missed. Asking 2 or 3 specific, standardized questions about SHS exposure, along with better cessation training, may help alleviate these barriers.
Caregivers may also be more likely to conceal their smoking status when they perceive a high smoker-related stigma,27 and the stigma may in fact be heightened for parents of young children,28 especially when discussing their smoking status with a health care provider. This heightened perceived smoking stigma may influence caregivers to conceal the details of SHS exposure when asked during admission screenings, although they may feel more comfortable with a relatively anonymous study team member. For example, the caregiver may assume that his or her smoking behavior outside of the home is not enough to harm the child and, therefore, does not need to be disclosed, especially given that many parents underestimate the harm that SHS has on children.29
Children who are admitted with respiratory illness were identified more often and more accurately. This finding may be because providers view SHS exposure as more clinically relevant and were inclined to ask questions about exposure. However, many of the health consequences of SHS are not related to respiratory illness6–9; this discrepancy represents a significant missed opportunity for intervention. Children who were exposed to smoke inside of the home were also more often and accurately identified by residents’ and nurses’ documentation. Those who are exposed in the home have the highest cotinine levels and the greatest exposure, and thus providers may be more likely to smell cigarette smoke in the patient’s room and be prompted to ask more questions. However, children exposed at lower levels are still at risk for negative effects,10,11 and the current findings reveal that many exposures also occur outside of the home. Our finding that the median cotinine level for those children who were exposed outside of the home was >1 ng/mL is consistent with a previous study that found the average cotinine level when the parent smoked outside only was 2.32 ng/mL.30
Finally, previous studies have cited time as a barrier to smoking cessation counseling.26,31 The most detailed method of assessing SHS exposure in the current study was the structured caregiver interview, in which caregivers were asked questions about their own smoking status and other sources of SHS exposure. They were also assured by the study team that their responses would be kept confidential. Thus, caregivers may have disclosed more details about their child’s SHS exposure than they would during a brief admission screening. Given the emotional stress on families during the admission process, as well as the time constraints on providers, it may be helpful to develop a 2-step screening process for pediatric inpatients: a quick but highly sensitive question or biological marker to identify any potentially exposed child, followed by a more in-depth interview with a cessation counselor to determine the extent of exposure and provide evidence-based cessation intervention.
There are a number of limitations to the current study. First, our response rate, even after the addition of an incentive, was low; this low rate may have introduced response or verification bias into the sample, and we do not have information on the nonrespondents. We were also unable to collect sufficient saliva to obtain cotinine levels for 59 patients, which resulted in small cell sizes for some analyses and limits our ability to make generalizations about the cotinine levels.
Because we are relying on chart review, it is possible that the providers did ask about SHS exposure and did not record the response in the chart, especially in cases in which no SHS exposure was present and thus was not deemed “relevant.” However, this possibility would not explain the high number of patients who were misclassified as not exposed to SHS, when they were exposed. Finally, this study represents the experience at one academic institution and may not be generalizable to other institutions, especially freestanding children’s hospitals or community hospitals.
Tobacco smoke exposure, even at low levels, is harmful for children. Screening children for SHS exposure needs to be done consistently and accurately in the inpatient setting, where parents may be more receptive to hearing about strategies to reduce SHS exposure for their children, as well as the importance of quitting smoking themselves. In this study, most children exposed to SHS were not identified as exposed according to documentation of admission screenings. Children admitted for conditions other than respiratory illness, and children exposed to SHS outside the home, seem to be the most vulnerable for being missed in admission screenings. Future research is important to identify accurate and efficient methods of screening for and accurately identifying SHS exposure among children admitted to the hospital. In the meantime, training providers to ask about all sources of SHS exposure, document those exposures, and arrange intervention for children with exposure will provide many more opportunities to protect children from the toxic effects of tobacco smoke.
The majority of smoke-exposed children in this study were not identified as exposed on the basis of documentation of admission screenings. Future research is important to detect accurate and efficient methods of screening for and identifying SHS exposure among children admitted to the hospital.
RESEARCH SUPPORT: This study was funded by the American Academy of Pediatrics/Julius B. Richmond Center of Excellence, through a grant from the Flight Attendant Medical Research Institute.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- confidence interval
- emergency department
- Golisano Children's Hospital
- secondhand smoke
- Teach SJ,
- Crain EF,
- Quint DM,
- Hylan ML,
- Joseph JG
- Weitzman M,
- Gortmaker S,
- Walker DK,
- Sobol A
- Weitzman M,
- Cook S,
- Auinger P,
- et al
- Chang JS,
- Selvin S,
- Metayer C,
- Crouse V,
- Golembesky A,
- Buffler PA
- Wilson KM,
- Finkelstein JN,
- Blumkin AK,
- Best D,
- Klein JD
- Winickoff JP,
- Hillis VJ,
- Palfrey JS,
- Perrin JM,
- Rigotti NA
- Irvine L,
- Crombie IK,
- Clark RA,
- et al
- Willers S,
- Axmon A,
- Feyerabend C,
- Nielsen J,
- Skarping G,
- Skerfving S
- Bernert JT Jr.,
- McGuffey JE,
- Morrison MA,
- Pirkle JL
- Benowitz NL
- Halterman JS,
- Borrelli B,
- Tremblay P,
- et al
- 25.↵Statistical Package for Social Sciences [computer program]. Version 15.0. Chicago, IL: IBM; 2006.
- Tong EK,
- Strouse R,
- Hall J,
- Kovac M,
- Schroeder SA
- Stuber J,
- Galea S
- Matt GE,
- Quintana PJ,
- Hovell MF,
- et al
- Copyright © 2012 by the American Academy of Pediatrics