Implementing Parental Tobacco Dependence Treatment Within Bronchiolitis QI Collaboratives

The BQIP and the SIB collaboratives were sponsored by the VIP Network, which is a QI network led by pediatric hospitalists. Both collaboratives were 1-year projects centered in a virtual community that shared information through webinars and a change package of evidence-based resources, online data sharing, real-time performance feedback, a project Listserv, and QI coaching. Twenty-one hospitals participated in the BQIP, and 35 hospitals participated in the SIB project; in both collaboratives, hospitals were geographically spread across the continental United States (Table 1). The preexisting climate around the screening and treatment of tobacco dependence in both collaboratives was assessed by self-reports from each site at the start of each project, including screening for TSE and the provision of smoking cessation interventions (Table 1). In the SIB collaborative, the specific intervention of recommending or providing NRT was also assessed. Additional methodologic details for the BQIP and SIB collaboratives are published separately.^14,18 

In both the BQIP and SIB collaboratives, each site was required to establish a project team that included a hospitalist and either a respiratory therapist or nurse, although teams were generally much larger. The SIB project included a representative from the emergency department. The project teams participated in webinars, including an hourlong session devoted to implementing tobacco-dependence interventions, during both collaboratives. There were scheduled conference calls throughout the collaborative and access to coaching from QI and tobacco control experts. All teams received a tobacco change package of interventions (Fig 1), including suggested best practices to increase the screening of children for TSE and provision of tobacco-dependence treatment and referrals for parents and/or caregivers. Although coaching was a part of the overarching QI project strategy, additionally, optional tobacco-specific coaching by phone and e-mail was offered to the sites, although we did not specifically track the use of this resource. Sites were able to access their performance in real time and received frequent feedback on their performance compared with the entire collaborative. Although many of the resources were educational in nature, our major focus in the projects was standardizing systematic screening for TSE and integrating tobacco-dependence interventions or referral into routine practice. The resources for both projects were made available on the VIP Network Web site after completion.^14,18 

In the BQIP collaborative, the tobacco-dependence treatment interventions recommended were based on the US Department of Health and Human Services (HHS) Clinical Practice Guideline 5 A’s approach.²⁰ In 2013, the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition combined the terms “abuse” and “dependence” for substances of abuse and used the nomenclature “tobacco use disorder.”²¹ To maintain consistency with the HHS guidelines and the AAP Tobacco Policy Statements, the term “tobacco dependence” is used in this article to denote parents and/or caregivers who use tobacco.^17,20 The HHS Clinical Practice Guideline recommends practitioners to do the following: ask about tobacco use, advise the tobacco user to quit with a personalized message, assess readiness to quit, assist with evidence-based therapies, and arrange follow-up (Fig 1).²⁰ A 3-step version was also provided, which has been used in pediatric settings: 2 A’s and 1 R (ask, advise, and refer).²² The recommended screening question for TSE was, “Does your child live with anyone who smokes cigarettes or other tobacco products?” It was suggested to add this question to the electronic health record in the nursing intake and physician templates to improve the reliability of screening.

On the basis of the lessons learned from the BQIP, the aims and interventions were more clearly specified in the SIB collaborative. For the SIB project, the tobacco-dependence treatment measures were further defined by using the Clinical Effort Against Secondhand Smoke Exposure model of the updated 2 A’s and 1 R (Fig 1).²³ The Clinical Effort Against Secondhand Smoke Exposure model includes both counseling and pharmacotherapy, which has been found to be the most effective in addressing tobacco dependence within a framework for busy clinicians and takes ≤3 minutes to complete.²⁰ The tobacco interventions recommended included personalized advice to quit smoking, referral to local resources or the state quitline, and recommendations or a prescription for NRT. We also recommended that sites add automatic tobacco-dependence interventions in their bronchiolitis order sets for children have positive screen results for TSE.

Tobacco-related measures for bronchiolitis patients were created by expert planning group consensus based on published national guidelines for adults and literature review for evidence-based tobacco-dependence interventions for parents and/or caregivers who smoke.²⁰ Screening for TSE and intervention rates were chosen as measures for both collaboratives. Screening was defined as the rate of documentation (positive or negative) of TSE in the chart over the total number of charts screened. The intervention rate was defined as the rate of documentation of cessation counseling or referral for services in the chart for children with positive screen results for TSE. In the SIB project, the provision of or recommendation for NRT was also defined on the basis of documentation of the action in the medical record.

For both collaboratives, data collection was performed by a manual chart review of the first 20 charts per month that met inclusion and exclusion criteria at each site for 2 subsequent bronchiolitis seasons, with the intervention period occurring between seasons. In the BQIP collaborative, preintervention data were collected from January 2013 to March 2013, and postintervention data were collected from January 2014 to March 2014; for the SIB project, preintervention data were collected from December 2014 to March 2015, and postintervention data were collected from December 2015 to March 2016. Data were entered by using the AAP Quality Improvement Data Aggregator, which is a centralized, online data repository. The Quality Improvement Data Aggregator provided real-time data on the project Web site, where researchers could compare local performance against group performance using run charts.

For the BQIP collaborative, the data were aggregated on a monthly basis and analyzed by using analysis of means (ANOM), which is a QI statistical method in which researchers compare each cycle mean to 3 σ control limits that are derived from the group mean to determine special cause variation.^24,25 Because the BQIP was our first QI collaborative and took place primarily in relatively low-resource community hospital settings, data were requested monthly to minimize the burden of data collection. The 6 months of available data did not allow for traditional statistical process control (SPC) run charts; thus, ANOM was used. On the basis of the success of the BQIP, for the subsequent SIB collaborative, we chose to collect weekly data and use SPC run charts to determine the presence of special cause variation using established rules.²⁶ In both projects, only those patients who had positive screen results for TSE (ie, parent or caregiver who smoked) were included in the denominator for further interventions, such as offering counseling, referral to resources, and recommendation of NRT.

To address the question of whether the iterative changes made in our approach to promoting tobacco cessation were linked to improved outcomes, we compared the results from each collaborative using relative risks (RRs) with the first collaborative (BQIP) as the reference.

The BQIP and SIB collaboratives were approved by the AAP Institutional Review Board as well as site institutional review boards as deemed necessary by the participating institutions. No protected health information was collected for either project.

For the BQIP collaborative, 21 hospitals participated and a total of 1869 charts were reviewed over the project: 995 in the preintervention period (January 2013 to March 2013) and 874 in the postintervention period (January 2014 to March 2014). In the BQIP collaborative, the preintervention mean rate of screening children for TSE was 79.5% and improved to 88.9% postintervention.¹⁴ Of the children screened, 416 (25%) were identified as exposed: 204 in the preintervention period and 212 in the postintervention period.

Of the identified parents and/or caregivers who smoke, 45 (22.1%) received either cessation counseling or referral in the preintervention period, and 109 (51.4%) did so in the postintervention period. Monthly performance improved from 18% to 26% during each individual month in the preintervention period to between 38% and 56% postintervention, as annotated in Fig 2. ANOM revealed special cause variation for the intervention and referral in the first 2 months postintervention although not in the final month.

For the SIB collaborative, 35 hospitals participated and a total of 4389 inpatient charts were reviewed. We reviewed 2202 charts in the preintervention period (December 2014 to March 2015), and 2187 charts were reviewed in the postintervention period (December 2015 to March 2016).

The preintervention collaborative mean screening of children for TSE was 80.6% and improved to 86.6%. Among tobacco-dependence interventions, there was an increase in counseling or referral to resources from 31% to 42% after the distribution of the change package (Fig 3). With continued coaching and feedback, the rate increased to 53% by the end of the collaborative (Fig 3). The rate of NRT recommendation or prescription rose to 12.3% after the distribution of the change package from 3.4% at baseline (Fig 4). Rules for special cause variation were satisfied for each metric, at which point the centerline was adjusted (Figs 3 and 4).

Because of the iterative nature of the collaborative, the outcome of screening for TSE and provision of interventions (counseling and referral) were compared between the BQIP and the SIB project by using RRs, as shown in Table 2. Despite the more intensive approach in the SIB collaborative, the relative improvements achieved in TSE screening and provision of smoking cessation counseling and referral were similar between the projects. In addition, the rate of children who had positive results for TSE was unchanged between the collaboratives. Because our aim in the recommendation or prescription of NRT was not specifically included in the BQIP, this measure was not compared.

In both collaboratives, we demonstrated an increase in the rates of interventions to address parent and/or caregiver tobacco dependence as 1 component of broader QI efforts. The novel component of this project was our focus on parental smoking cessation rather than other common pediatric methods for reducing child TSE, such as promoting indoor smoking bans. Although both collaboratives revealed substantive improvements in our chosen measures, the iterative changes between the successive collaboratives did not result in additional increases in screening or the rates of counseling and referral for the parents of tobacco-exposed patients. The rate of NRT referral and/or prescription in the SIB collaborative did increase by fourfold; however, the change between collaboratives could not be compared because we did not address NRT in the BQIP.

Studies to reduce parent and/or caregiver smoking in pediatric settings have revealed that more intensive efforts, including pharmacotherapy (eg, NRT), are more successful.^27,–29 However, adding a specific aim to recommend or prescribe NRT in the SIB project could have distracted sites from their efforts to perform counseling and referral. In addition, providers may have had decreased levels of comfort with recommending a medication for parents and/or caregivers, and this may have limited the overall interest in providing tobacco-dependence treatment. However, it should be noted that increasing the tobacco-related intervention aims from 1 measure in the BQIP to 3 measures in the SIB collaborative did not decrease the rates of improvement.

Previous studies in which researchers implemented systems-change interventions for smoking cessation revealed challenges because of perceived barriers by health care providers, including lack of time, skills, training, and resources.³⁰ We addressed these previously reported barriers in our tobacco change package, which is supported by group training and individualized coaching. There are several smaller studies in the pediatric inpatient setting that revealed the acceptability and feasibility of implementing interventions for parent and/or caregiver tobacco dependence.^31,–34 Several pediatric studies also include NRT as a component of interventions for parent and/or caregiver tobacco-dependence treatment.^22,27,31,33 Similar to public health recommendations for prescribing prophylaxis to families for exposure to infectious diseases, such as pertussis and meningococcemia,³⁵ pediatricians can prescribe or recommend NRT for a tobacco-dependent parent and/or caregiver who is interested in treatment.

We believe there is an ethical imperative to address parental tobacco dependence with evidence-based therapies to improve child health, particularly among infants who have been hospitalized with bronchiolitis because there is overwhelming evidence of the relationship to TSE.^3,19 The Surgeon General’s Report revealed that there is no safe level of TSE, and the most effective method to eliminate a child’s exposure is smoking cessation.¹ Thus, the interventions in the change package were focused on tobacco-dependence treatment rather than harm reduction. In the United States, 68% of adult smokers want to quit smoking, but the vast majority do not receive meaningful assistance; thus, pediatricians are increasingly addressing this unmet need.³⁶ 

Children who have parents and family members who smoke cigarettes are more likely to initiate smoking, perpetuating the cycle of economic and health disparities.^37,–40 Parents of young children are often more likely to see their children’s physician than their own (if they have 1), and pediatricians frequently provide treatment and recommendations for family members to address pediatric issues ranging from maternal postpartum depression to household dietary and activity changes.^41,42 Thus, in these collaboratives, we sought to capitalize on a classic teachable moment to deliver health promotion advice to parents. Additionally, we sought to create a teachable moment for pediatric clinicians to think differently about their approaches to parents who smoke as a secondary benefit of participation in a QI collaborative.

There are multiple limitations to this analysis, including the potential for discrepancy between the documentation of provision of tobacco-dependence interventions and the actual delivery of such interventions. Chart documentation does not always correlate with performance of the action of interest, although it is a reasonable proxy. In addition, there can also be discrepancy between reported and actual TSE. Wilson et al⁴³ demonstrated that parental and caregiver report in the inpatient setting often misses children with TSE when biochemically confirmed. Furthermore, we only addressed process measures around tobacco use and did not collect data on quit attempts or objectively measure any decrease in child TSE. Despite the limitations inherent in our approach, if the systematic implementation of tobacco-dependence treatment of parents was added as an inpatient quality measure for bronchiolitis, the ultimate public health impact could be significant.

Although rates of TSE screening are generally high, future researchers in this area should prioritize methods to further promote cessation counseling and referral because despite major improvements, approximately half of all smoking parents identified in our study still did not receive a cessation message. Furthermore, studies will be needed to clarify the appropriate intensity of tobacco-dependence training and interventions in the hospital setting. Finally, studies are needed to confirm that QI interventions truly correlate with increased quit rates among parents and/or caregivers and/or can produce biochemical confirmation of a reduction in child TSE.

Interventions to promote tobacco-dependence treatment of parents and/or caregivers were successfully implemented as a component in 2 consecutive QI collaboratives with the primary goal of improving care for children who are hospitalized with bronchiolitis. The appropriate intensity of the training and intervention strategies requires further study.

AAP

American Academy of Pediatrics

ANOM

analysis of means

BQIP

Bronchiolitis Quality Improvement Project

HHS

Department of Health and Human Services

NRT

nicotine replacement therapy

QI

quality improvement

RR

relative risk

SIB

Stewardship in Bronchiolitis

SPC

statistical process control

TSE

tobacco smoke exposure

VIP

Value in Inpatient Pediatrics