Study Setting {9}
The New York City Housing Authority (NYCHA) is the largest public housing provider in the US with over 500,000 residents across 335 developments throughout the city via conventional public housing, Section 8, and Permanent Affordability Commitment Together (PACT) or Rental Assistance Demonstration (RAD) housing (64). Though NYCHA is unique in its size and organization of buildings, it is characteristic of other public housing settings in that its entire portfolio of buildings and campuses are subject to smoking bans. Additionally, the resident profiles are similar across socioeconomic status and smoking behaviors. For feasibility, this study will be conducted at public housing sites located across two New York City boroughs where public housing is most concentrated: Manhattan and the Bronx. There are a total of 159 developments and 93,012 current dwelling apartments in these two boroughs. Of these, we have restricted the pool of buildings to contain only buildings managed by NYCHA. Further building inclusion/exclusion criteria are detailed below.
Eligibility Criteria {10}
There are two sets of eligibility criteria: (1) building eligibility and (2) participant eligibility.
The first stage of enrollment occurs at the building level as the buildings are randomized to receive a designated intervention (A, B, A + B) or act as a control site. Buildings must (1) have more than 50 units and (2) not be undergoing major renovations. Buildings that belong to developments that are (1) not in Manhattan or the Bronx, (2), smaller than 50 units, (3) undergoing major renovations, (3) mixed finance, (4) exclusively for elderly, (5) privately managed, (6) or will be part of Rental Assistance Demonstration (RAD) or Permanent Affordability Commitment Together (PACT) are excluded. RAD and PACT programs are associated with increased funds towards repairs and investments in social programming that have the potential to influence the impact of the designed interventions (65, 66), and, thus, have been excluded (44).
Upon randomization, resident leaders will be informed of the study on a rolling basis. If a resident leader declines participation of their buildings and residents in the study, their buildings will be removed from the pool of eligible buildings. See Fig. 3 for the site selection process.
The second stage of enrollment is at the resident level. Participants must be (1) able and willing to provide verbal informed consent, (2) at least 18 years old, (3) able to communicate in English or Spanish, (4) living in the building at least 5 days/week and 9 months/year, and (5) not planning to move in the next 2 years. Participants who (1) have severe physical or mental medical conditions (e.g., cognitive disability) or other factors that could limit participation or ability to give informed consent in the study at baseline or during follow-up visits, (2) participate in focus groups conducted pre or concurrently as RCT that are geared towards understanding how to improve intervention arms for greater impact (buildings identified for focus groups are buildings that will not be assigned an intervention or have recruitment conducted for the main study), or (3) only smoke non-tobacco products (e.g., marijuana) are excluded from participating in the study due to the focus of the study on tobacco products. It should be noted that during the period of recruitment, the smoke-free policy in NYCHA was modified to also include marijuana use. In anticipation of this modification to the smoke-free policy, we are collecting preliminary information regarding marijuana use.
Recruitment of NYCHA residents will be conducted via door knocking and lobby intercepts until the targeted number per group is reached (4 smokers, 4 non-smokers in each building). Smoker status is defined as smoking a tobacco product including e-cigarettes at least 5x/month, and nonsmoker status is defined as never having smoked any product or quit smoking at least 12 months prior to recruitment.
Who Will Take Informed Consent? {26a}
Participant contact will be done face-to-face with systematic questionnaires and potentially identifying information will be asked or recorded for study participants (although NO birthdates, photos, or video recordings will be taken of participants) by trained field research support staff. Consent to participate in the study will be asked before the actual study begins and individuals who refuse will not take part in the study. We will explain the study, its intent, its potential risks, and ask if we can proceed. Informed consent documents and study materials will be translated (forward and back translated) and administered in Latin Spanish where needed. Any information that is collected and is recorded first on paper will then be entered into a computer and transferred and stored in a secure electronic environment. Upon conclusion of the study, all identifying information for each participant will be destroyed.
Additionally, a waiver of written documentation of informed consent has been obtained through the overseeing IRB as the research presents no more than minimal risk of harm to subjects and does not involve procedures for which written consent is normally required outside of the research context. A waiver of documentation of consent is appropriate because the only link between the subject and the study would be the consent document and the primary risk is a breach of confidentiality.
Additional Consent Provisions For Collection And Use Of Participant Data And Biological Specimens {26b}
The study data will be released to the funder, the National Institutes of Health (NIH) and National Cancer Institute (NCI) at the conclusion of the study. Identifiers will be removed and, after such removal, the information or biospecimens could be used for future research studies or distributed to another investigator for future research studies without additional informed consent, as appropriate.
Interventions
Explanation For The Choice Of Comparators {6b}
Buildings and study participants assigned to the comparator arm [Arm 4] are recruited and followed over approximately a 12-month period to assess outcomes. No additional programs or services will be delivered to the buildings or residents assigned to this arm beyond standard programs that NYCHA may provide to support the smoke-free mandate. Field staff will document any policy-related signage, activities, or information to which these participants are exposed. We expect and have confirmed that Smoke-Free NYCHA’s liaison program will occur during our period of data collection during which they will conduct programming and activities that support compliance to the smoke-free policy. As such, the buildings enrolled in our ‘Standard NYCHA approach’ are buildings that the Smoke-Free NYCHA team has identified as locations for concentrated efforts of their own smoke-free programming over the course of our data collection period. We will therefore evaluate the impact of this parallel prevention program including a covariate for NYCHA study participation at baseline in the models specified below. We will further allow for an interaction between the NYCHA study participation indicator and the treatment group.
Intervention Description {11a}
The development of the interventions is grounded on prior smoke-free policy endorsement and tobacco reduction/cessation programs. NYCHA conducted a smoke-free pilot project in 2015 centered on a smoke-free home pledge at a housing complex in Upper Manhattan (67). The initiative was organized by members of the tenant association and entailed families and neighbors taking a voluntary pledge to maintain a smoke-free living environment. The overwhelming majority of residents (85%) signed the pledge, including nine floors in which all apartments took the pledge. At a culminating celebration, residents shared personal accounts of motivations for endorsing a smoke-free living environment (67). Several elements of the “resident endorsement” intervention arm are fashioned after this pilot initiative, including fostering social connections between residents, providing information on the hazards of smoking and secondhand smoke along with the pledge culminating in a town hall session.
For the “reduction/cessation” intervention arm, Dr. David Albert, DDS, has led and directed the design of the tobacco cessation protocol. He also provided the foundational tools and trainings necessary to effectively use motivational interviewing to not only support smoking cessation, but also to reduce personal smoking. Dr. Albert’s expertise comes from implementing and evaluating evidence-based tobacco cessation programs at the New York Presbyterian (NYP) Cornell and NYP Columbia campuses and within affiliated community-based organizations, including many that engage minority populations. At NYP Hospital he is the Tobacco Cessation Project Lead for the Medicaid Redesign Project funded by the NYDOH Delivery System Reform Incentive Payment Program. This program trains clinicians through a certified tobacco cessation specialist program, implements tobacco cessation counseling and offers community presentations and engagement. As the Co-PI for the NCI funded Tobacco Cessation via Public Health Dental Clinics, he has examined the effectiveness of brief office- based interventions designed to assist patients quitting smoking or smokeless tobacco use. Dr. Albert is the co- Project Lead of the Development of a Tobacco Cessation Program for Cancer Patients at Columbia University. The project aims to implement an evidence-based tobacco cessation program for smoking patients at the Columbia University Herbert Irving Comprehensive Cancer Center.
To further enhance the intervention curriculum, we have conducted key informant interviews with residents and experts in resident engagement, tobacco cessation and housing-based health interventions. These interviews were designed to elicit strategies to not only improve the RCT approach, but also to ascertain necessary adaptations to the intervention protocols to maximize their efficacy. Such discussions led to expanding the scope of the reduction/cessation intervention which previously was designed to focus on relocating smoking and cessation only. Furthermore, our community partner, Health People have previously conducted health education activities in similar settings and their expertise in resident facing activities will further enhance resident engagement.
Prior to the implementation of the intervention, recruitment and baseline data collection via survey questionnaire, passive drool sample collection and sensory building observations will be conducted. Trained study staff will obtain verbal consent upon the time of recruitment and data collection prior to enrollment in the study. Approximately a month after this completion, interventions will be implemented by peer educators in accordance with respective building assignments.
Arm 1: Reduction (via relocation and reduction in personal smoking)/Cessation
The reduction/cessation arm (Arm 1) “meets smokers where they are” using a) a harm reduction approach that reinforces the use of designated smoking areas available and/or identifies ways residents can reduce their personal smoking more generally via motivational interviewing skills thereby encouraging compliance via relocation and personal smoking reduction; along with b) in-residence cessation support to reduce barriers to participation in cessation services. Enrolled participants who are smokers will be referred by the survey team to peer educators from Health People. The peer educator will coordinate smoking cessation support, including serving as a liaison between participant and research team, providing information regarding the smoke-free policy and opportunities for relocation, and connecting participant to tobacco replacement therapy and/or physician support if deemed appropriate.
Arm 2: Resident Endorsement
The resident endorsement model (Arm 2) seeks to engage residents by using an empowerment/network-based approach to shift the culture of health in buildings that encourages residents to collectively tackle indoor smoking and secondhand smoke exposure via physical demarcations and/or social media. Buildings assigned to this arm will be targeted for a series of 2 in-residence programs that involve community forums and the creative arts to garner resident endorsements of smoke-free living environments. Premised on resident engagement, this arm seeks to impact social and physical dimensions of the residential environment to achieve compliance. The sessions will: 1) inform residents of risks associated with smoking and secondhand smoke; 2) identify reasons to have a smoke-free home, 3) ask residents to sign a pledge on paper and/or virtually; 4) display smoke-free signage on doors and/or social media pages with an original hashtag (#Smokefree[building address]); and 5) refer residents to the Smoke-free NYCHA website for information on the policy and existing cessation resources.
Arm 3: Combined Intervention
There will be buildings assigned to receive both models (Arm 3) for which in-residence programs based on the resident endorsement treatment and the smoking reduction/cessation treatment will be provided. Both will occur simultaneously with one geared toward all building residents (resident endorsement) and the other targeting smokers (smoking reduction/cessation) with the goal of reducing both personal smoking and secondhand smoke exposure.
The interventions and comparison group are outlined in Table 1.
Criteria For Discontinuing Or Modifying Allocated Interventions {11b}
The criteria for discontinuing or modifying allocated interventions in any given building will be in response to any requests by overseeing building authority/management who may voice concerns about administering interventions and decline further participation of their building, and subsequently residents, in the study.
The criteria for discontinuing targeted intervention approaches like the reduction/cessation and combined interventions will be in response to participant request to discontinue or nonresponse.
Strategies To Improve Adherence To Interventions {11c}
Interventions are administered by peer educators who will collect evaluations to assess quality of interactions and sessions. These are reviewed by researchers periodically to ensure participants are engaging with the activities in a meaningful way and appropriate protocols in delivering the interventions are followed. Research team members will also periodically attend sessions related to resident endorsement activities and meetings with participants related to the reduction/cessation activities to further ensure appropriate protocols are taken.
Relevant Concomitant Care Permitted Or Prohibited During The Trial {11d}
Participants will not be restricted from pursuing additional/further tobacco reduction/cessation support services. We will make every effort to document any services provided by NYCHA during regular briefings, and also in assessments.
Provisions For Post-trial Care {30}
For the reduction/cessation intervention, individuals interested in quitting will be provided a month supply of nicotine patches. There are limited risks and steps are taken to ensure those more likely to experience possible side effects are not provided patches. Peer educators will report any instance of harm to the Columbia research team who will then take action to remedy the situation as deemed appropriate and report the adverse event accordingly.
Outcomes {12}
Primary Outcomes:
There are three primary outcomes being measured:
- We will evaluate the change in number of cigarettes smoked per day using self-reported average number of cigarettes smoked per day among smokers measured at the time of baseline assessment, interim assessment (approximately 3 months post-intervention), and long-term assessment (approximately 12 months post intervention).
- We will evaluate change in salivary cotinine levels for 25% of the sample via passive drool collection among smokers and non-smokers collected at baseline assessment and the long-term assessment.
- We will evaluate change in secondhand smoke exposure via self-reported amount of exposure to secondhand smoke. This will be measured in all three assessments.
Secondary outcomes
There are 6 secondary outcomes being measured:
- We will evaluate change in number of participants with successful quit attempts measured at baseline assessment, interim assessment, and long-term assessment for smokers.
- We will evaluate change in number of quit attempts using the mean number of quitting attempts among smokers measured at baseline, interim, and long-term assessments.
- We will evaluate change in number of participants with second-hand smoke observations who have either observed second-hand smoking or not measured at baseline, interim, and long-term assessments.
- We will evaluate change in number of hours of second-hand smoke exposure via self-reported number of hours of observing someone smoke indoors measured at baseline, interim, and long-term assessments.
- We will evaluate change in number of smokers via counted number of people observed smoking in common areas at building visits.
- We will evaluate change in number of cigarette butts via counted number of cigarette butts observed in common areas at building visits.
These outcome measures are summarized in Figure 5.
Participant Timeline {13}
Consented, eligible participants will be in the study for up to 15 months once recruited. A staggered recruitment, data collection, and intervention delivery model will be implemented to ensure appropriate timing of participant involvement from recruitment to the final follow-up and to better align study resources across all 64 enrolled buildings (See Table 2.)
Sample Size {14}
Power analysis was conducted using software PASS.15 for the primary outcome (change in salivary cotinine over a year) in Aims 1 and 2. Sampling 16 clusters (buildings) with 4 subjects per cluster (smokers for Aim 1, non-smokers for Aim 2) in each arm, we have 128 subjects per group or 256 in groups with or without a specific type of treatment (e.g. resident endorsement vs. no endorsement, cessation vs. no cessation). We calculated effect size as group mean difference in unit of standard deviation (SD) of the outcome, based on a two-sided test for mean difference between two groups in a cluster-randomized design with power of 80%. To compare groups with and without specific type of treatment, the group size of 256 may detect effect size of 0.283SD, 0.299SD, 0.312SD at significance level alpha = 0.05 for intra-cluster correlation coefficient r = 0.1, 0.15, 0.2, respectively. To compare intervention arms to the control arm, we have 3 comparisons and use conservative Bonferroni adjustment for multiple tests on significance level that alpha = 0.0167 (= 0.05/3). The sample size of 128 subject per arm may detect effect size of 0.472SD, 0.499SD and 0.524SD for r = 0.1, 0.15 and 0.2, respectively, at alpha = 0.0167. To put these numbers in perspective, using mean and SD estimated from a study of smokers in Maryland, an effect size of 0.283SD would correspond to a mean difference in salivary cotinine between the two intervention groups of 55.3 ng/mL with group SD = 195.4 ng/mL for smokers and of 0.19 ng/mL with group SD = 0.672 ng/mL for non-smokers.124 For instance a mean difference in 5 cigarettes smoked per day in the same study was 59.6 ng/mL for smokers.124 In England, following the smoke-free legislation salivary cotinine declined from mean level of 0.36 to 0.07 ng/mL (mean difference 0.29 ng/mL) among non-smokers.125 These findings support that we have sufficient power to estimate the impact of our interventions. The study is powered on an 80% retention rate, which we have achieved in other studies. Reasons for attrition will be tracked; those lost to follow up will be compared to the final sample.
Recruitment {15}
The primary method of recruitment will be via door knocking and lobby intercepts until we reach our targeted number of residents per group (4 smokers, 4 non-smokers in each building). We will make every attempt to retain participants and minimize attrition bias through a range of rapport building and retention techniques, incentives, and cultural adaptation of study materials. We have also powered the study based on 80% retention.
While the study presents minimal risks, participants may withdraw at any time upon request.
Alternatives available to participants outside the research context are simply not enrolling and attempting to serve the greater good by participating in other research or community service activities.
Assignment of interventions: allocation
Sequence Generation {16a}
A list of eligible buildings from each borough will be randomly ordered by computer-generated random numbers for each borough. Then, the first 24 buildings for each borough (n = 48, 24 per borough) will be randomly assigned to receive intervention A, B, or A + B.
Additionally, all buildings that are eligible for the study but have been identified by NYCHA as locations where they will conduct activities to promote the smoke-free policy will be randomly ordered by computer-generated numbers for each borough. Then, the first 8 buildings for each borough (n = 16, 8 per borough) will be assigned to serve as a control site.
Concealment Mechanism {16b}
The allocation of interventions is done on a building basis and will only be revealed to specific research team members and intervention implementors (PEs). This information will be passed to Health People via encrypted emails. All data collectors will not be informed of intervention assignment for the duration of the study.
Implementation {16c}
Dr. Valeri oversaw the allocation sequence and intervention assignment prior to the start of recruitment. Trained field-based research support staff are recruiting and enrolling participants in buildings that are enrolled in the study.
Assignment of interventions: Blinding
Who Will Be Blinded {17a}
Trial participants and outcome assessors are blinded after the assignment to interventions. All interventions are assigned at the building level. We have a separate field team research support staff who conduct recruitment and collect the assessments but are unaware of the assignments.
Procedure For Unblinding If Needed {17b}
Overall, the study has very minimal risks, and by its nature participants may know which intervention they/their building is receiving based on what services, if any, are provided and what materials are displayed, if any. Blinded data collectors may also guess the intervention based on what materials are displayed or not displayed. The intervention team, the peer educators will report any serious adverse events that occur during intervention implementation to the research team, but the identity of the participant and the intervention assigned to their building will not be revealed to data collectors.
Data collection and management
Plans For Assessment And Collection Of Outcomes {18a}
There are three assessments during this trial. These include a baseline assessment conducted upon recruitment and after building randomization, an interim assessment conducted over the phone approximately 3 months after the delivery of intervention, and a final long-term assessment conducted approximately 12 months after the delivery of intervention. The survey questionnaire collects information on housing conditions, indoor air quality, personal smoking behavior and cessation attempts, exposure to secondhand smoke, opinions on secondhand tobacco exposure, the smoke-free policy and tobacco policies, and health conditions. Additionally, passive drool samples from 25% of participants (n = 128) is collected at the time of the baseline assessment and at the final follow-up. Sensory building observations occur at the start of recruitment and during the final assessments to record signs of active smoking and/or tobacco/marijuana odor. The presence of cigarettes, cigars, ashtrays with/out ashes, matches or lighters will be documented within apartments and around the building premises including stairwells, elevators, hallways, laundry rooms, in/outdoor community spaces, playgrounds, and the perimeter of the buildings.
Plans To Promote Participant Retention And Complete Follow-up {18b}
We make every attempt to retain participants and minimize attrition bias through a range of rapport building and retention techniques, incentives, and cultural adaptation of study materials. Though primary outreach is conducted in-person, we also conduct mail and text outreach to maximize participant engagement. We have also powered the study based on 80% retention.
While the study presents minimal risks, participants may withdraw at any time upon request.
Alternatives available to participants outside the research context are simply not enrolling and attempting to serve the greater good by participating in other research or community service activities.
Data Management {19}
We will use a digital data capture system and collect data on password-protected laptop computers. Specifically, this multi-user system is registered with CUIMC IT. Data will be transferred, encrypted, backed-up and transferred to the data manager who will review the forms daily for completeness.
Data will be transferred via a firewall-protected, secure, electronic file-transfer to computer server space dedicated to the proposed study and protected with firewall and encryption technologies. Only Dr. Hernández, the Project PI, will have master access to this server and she will be required to authenticate herself via password each time the data are accessed.
Because this study is funded by the National Institutes of Health (NIH), the study is automatically issued a federal Certificate of Confidentiality. It is possible that participants will disclose information during their participation in the study and/or interviews and focus groups that indicates their involvement in illegal activities or indicates illegal activities by others. A Certificate of Confidentiality is one further means of ensuring that this information is kept confidential. Should participants disclose information in the middle of an interview that clearly indicates that they or those around them are in imminent danger, the interview will be suspended and 9-1-1 will be contacted.
Personally identifiable information (PII) will be collected and entered into an electronic database for each study participant. This PII data will be necessary to the conduct of the proposed trial and include: name; street address, city, county, and zip code; telephone numbers; and possibly electronic mail addresses. All data will be specifically used for the purposes of the proposed research. No PII will be disclosed to anyone who is outside of the proposed research team and all PHI will be destroyed when the study is concluded. We will institute strict precautions and security procedures to maintain data integrity and confidentiality:
- All data will be stored on a centralized computer server at the Columbia University Mailman School of Public Health. This server will be dedicated to the proposed study and will be protected with up-to- date "fire-wall" electronic security technology.
- Only Dr. Hernández, the Project PI, and the Project Coordinator will have master access to this server and the data.
- Computer administrators, co-investigators, a project manager, project coordinators, and laboratory staff will only be granted data access at the discretion of the Principal Investigator. Furthermore, this access will only be granted as needed for finite periods of time throughout the study.
- On a regular daily basis, laboratory staff will destroy any "portable" data (including both electronic and paper copies) that are transferred from external data sources to the university computer server.
- Once all records are linked and compiled into a single database, any and all PII identifiers will be destroyed.
The PI and the project coordinator will be responsible for providing and documenting appropriate user access to the study database and preventing against major sources of data security problems: unauthorized internal access to data, external access to data, and malicious intent to destroy data and systems. This user access will ensure that only appropriate and authorized personnel are able to view, access, and modify trial data.
Modifications to data will be performed in a manner that documents data modification, user access associated with modification, data associated with modification, and values prior to modification. The PI and the project coordinator will also be responsible for optimizing database performance, reliability and backup of data. External, unauthorized access to data is prevented through cooperative efforts of the PI, the project manager, and network and systems administrators. Highly successful measures will be employed through network firewall technologies to prevent unauthorized external access to data repositories.
Confidentiality {27}
We will establish excellent rapport with respondents and assure them that we are researchers with a private, local university and in no way representing NYCHA, law enforcement, or government agencies. No analyses, reports, or peer-reviewed articles will identify any participants. In later phases of the study, in-depth interviews may also be conducted somewhere nearby but not in the respondent’s household, such as a porch, church, park, café, etc., yet regardless of location all protections and confidentiality protocols will apply.
We will record information about the participant during initial participant recruitment that will allow us to recontact that same individual for follow-up surveys at 3 and 12-months post- intervention. With the participant’s consent, this information may include address and/or telephone number so that we can recontact them to schedule subsequent follow-up, possibly over the telephone. Participants will also be informed that any other identifying information that is recorded – such as names, birthdates, etc. – for study participants, will be destroyed at the conclusion of the study.
Plans for collection, laboratory evaluation and storage of biological specimens for genetic or molecular analysis in this trial/future use {33}
Exposure to smoke either directly among smokers or indirectly through secondhand smoke will be assessed through self-report and salivary collection.
We will collect saliva for 25% of recruited individuals from each building on a rolling basis at baseline and at 12 months via passive drool collection. Donors tilt their head forward, allowing the saliva to pool on the floor of the mouth, then pass the saliva through the SalivaBio Collection Aid (SCA) into a polypropylene vial. Collection protocols/methods are available online at www.salimetrics.com or upon request (Salimetrics Cotinine ELISA Kits). Saliva cotinine will be measured at the Columbia University Biomarker Core Laboratory directed by Dr. Regina Santella using cotinine ELISA Kits (Salimetrics, State College, PA). The Salimetrics ELISA kit is an enzyme immunoassay used to measure primary or secondhand exposure to nicotine via cotinine. It is not intended for diagnostic use. It is intended only for research use among humans and some animals. The assay is non-invasive, involves no genetic sequencing, and is not used as a diagnostic procedure.
Statistical methods
Statistical Methods For Primary And Secondary Outcomes {20a}
We will conduct intent-to-treat (ITT) and contamination-adjusted intent-to-treat (CA-ITT) analyses to estimate the effects of the smoking relocation/cessation and resident endorsement interventions compared to the standard program implemented by NYCHA (no treatment). CA-ITT analyses will account for unanticipated contamination between trial arms and be completed via two-stage instrumental variables regressions using the original random allocation codes per assigned unit. Preliminary data analysis will include examination of distribution and summary statistics of all variables by intervention groups at each time point. The continuous variables with skewed distribution will be properly transformed, if necessary, to reduce impact of extreme values or improve model fitting. We will use box-plots to examine how distribution of a quantitative variable varies by categories of a categorical variable, and use scatter plots and Spearman correlation coefficient to examine bivariate associations between quantitative variables. Chi-square and Kruskal-Wallis tests will be used to detect group differences in categorical and quantitative baseline variables, respectively. For successful randomization, we expect no difference in the distribution of all baseline variables among the four arms. The baseline variables that differ by intervention groups and are related to outcomes of interest will be controlled in the aim specific models.
Aim Specific
We will use generalized linear models with repeated measures (GLMRM) for each aim where outcome variables may have different form (continuous, binary or count) and the measurements are likely to be correlated due to residing in the same building or being from the same participant over time. The models use various link functions to relate outcome to linear combination of predictors. For example, identity link for continuous outcome is linear model with repeated measures and logit link for binary outcome is logistic model with repeated measures. We will use the generalized estimation equation method to estimate model parameters and make statistical inference since it takes into account intra- cluster (within-building or within-person) correlations and uses all available data.
1) Primary outcome
Self-reported average number of cigarettes smoked per day and secondhand smoking exposure (hours of secondhand smoke exposure in the building in the past 7 days) will be measured at baseline (in person interview), 3 months (phone interview) and 12 months (in person interview). Salivary cotinine will be measured for 25% of the sample at baseline and at the 12-month follow-up for smokers to evaluate reliability of self-reported smoking behavior in Aim 1 and for non-smokers to evaluate self-reported secondhand smoke exposure in Aim 2.
To examine the intervention group differences in the change of self-reported smoking (Aim 1), we will use GLMRM with identity link, E(Yijkh) = ßh + ßkhtk + ßZij, where Yijkh is an outcome variable (transformed if necessary) for smoking exposure measured at time tk (t0 = 0 for baseline; t1 = 1 for 6 months, 0 else; and t2 = 1 for 1 year, 0 else) from the ith participant of the jth building in the hth arm/group (h = 1, 2, 3, 4). For the reference group of h = 4, ßk4 = 0 for k = 0, 1, 2; so that parameter ßkh (k = 1, 2; h = 1, 2, 3 for intervention arm) indicates the difference in outcome change since baseline between intervention group h and reference group, and ß is a vector of coefficients for the vector of control variables Zij (if any). To examine if effect of treatment A depends on treatment B or vice-versa, we will test null hypothesis ßk1 = ßk2 + ßk3 for all k, which suggests independent effects of treatment A and treatment B and simpler model that E(Yijk) = ß0 + ß1A + ß2B + ßk0tk + ßk1tk A + ßk2tk B + ßZij with dummy variables A and B indicating treatment type. To examine the intervention group differences in the change in hours of secondhand smoke exposure in the building in the past 7 days for non-smokers, we will use GLMRM with log link (Aim 2). Models will be run with and without control variables (if any).
1) Secondary outcome and Sensitivity analyses
We will use GLMRM with logit or log links for binary or count outcomes, respectively, to evaluate group differences in the outcome changes from baseline to 3 and 12-month follow-up. The models will include predictors of dummy variables for time and group assignment as well as for group by time interaction as described above. Aim 1 (smoking behavior) secondary outcomes are whether smokers have successfully quit (binary), the mean number of quitting attempts (continuous). Aim 2 (exposure to secondhand smoke) secondary outcomes include ever observing someone smoking indoors within the building in the past 7 days (both smoker and non-smoker participants) (binary), hours of secondhand smoke exposure in the building in the past 7 days (count), number of people smoking and number of cigarette butts in common areas of the building (count).
Differential effect of the tobacco retail environment. The impact of differential tobacco retail environments (e.g., retailer density) around each building will be systematically evaluated for the primary outcomes and for quantitative secondary outcomes listed in Aims 1 and 2. The tobacco retail environment will be measured at the longitude-latitude point coordinate of each building using a geographic kernel density estimate with standard bandwidth to estimate the square mileage density of tobacco sellers in the surrounding area for each specific building point coordinate. This information is publicly available from the City of New York. As continuous metrics, these kernel density estimates will be separated in tertiles and at median breaks and the relationships tested in Aims 1 and 2 will then be tested within separate levels of tobacco retail density. We will also extend the GLMRM with identity or log link for each outcome variable in Aims 1 and 2 by adding the factor for retail environment, its interaction with group, time, and group by time interaction. In addition to the analysis for effect of a pre-specified modifier (the retail environment), we will conduct several a priori subgroup analyses to explore potential differential effect of the interventions in the continuous outcomes by age, sex, knowledge, and attitude towards the policy.
We will evaluate reliability of self-reported measures of active and passive smoking and revisit the primary analyses adjusting for potential imperfect reliability using regression calibration approaches.
Interim Analyses {21b}
Interim analyses will be conducted periodically for quality assurance purposes. These interim analyses will be reviewed by the research team to verify data quality and assess preliminary results.
Methods For Additional Analyses (E.g. Subgroup Analyses) {20b}
Key Informant (KI) interviews with individuals involved in building management and maintenance staff and those with expertise in the field (n = 20) was conducted to understand the current state of smoking at NYCHA sites and elicit their ideas about why smoking may persist in a mandated smoke-free context. We probed for recommended changes in the physical and social environments of the building including designated smoking areas and the information and support needed to more effectively support the implementation of the smoke-free policy.
Focus groups with 6–8 residents will be recruited via door knocking and lobby intercepts in non-study sites (n = 3–8 groups) on an ongoing basis. Focus group participants (n = 24–64) will be separated into English and Spanish led groups and discussion will be conducted using a semi-structured interview guide to elicit their knowledge and perspectives of smoking and health in buildings and to provide recommendations and actionable steps to achieve greater compliance among fellow residents. Selecting non-study sites will reduce contamination risk and engage more residents in discussions about smoke-free housing.
Methods in analysis to handle protocol non-adherence and any statistical methods to handle missing data {20c}
Missing data, due to nonresponse or incomplete records, may lead to information biases. To account for the potential effects of missingness and data missing at random or with no known pattern, we will use multiple imputation techniques. Multiple imputation datasets of any missing data will be made under a joint model for the variable in question and a missingness indicator conditional on the fully observed data. Estimates obtained from the multiply imputed datasets will be combined using Rubin’s Rule (Little, R. J., & Rubin, D. B. (2019). Statistical analysis with missing data (Vol. 793). John Wiley & Sons.)
Plans To Give Access To The Full Protocol, Participant Level-data And Statistical Code {31c}
Data from approximately 512 participants will be released to sponsor (NIH) once the final dataset is analyzed and main findings are accepted for publication to support and validate study findings. This will include participant demographics, data from interviews, and laboratory data from saliva samples. Identifiers might be removed from the identifiable private information or identifiable biospecimens and that, after such removal, the information or biospecimens could be used for future research studies without additional consent from the subject. De-identified participant data may be utilized for the purposes of repeated analyses by other researchers to verify findings, and/or to promote further research with new or alternative hypotheses. The mechanism of distribution will be a data sharing agreement.
Access will be granted for researchers, institutions, and/or the broader public for as long as the data is anticipated to be useful. With the data sharing agreement in mind, the data can be made available for the duration of time needed to conduct analyses.
Access will be granted to those with a reputable background in the scientific field who have either a scientific or medical degree and/or a relevant position to ask for the data. Individuals should also express their intended use of the data. These requests will be routinely reviewed by the principal investigator [and/or designated team member].
Oversight and monitoring
Composition Of The Coordinating Centre And Trial Steering Committee {5d}
The composition of the coordinating team and relevant partners are depicted in Fig. 6.
Composition Of The Data Monitoring Committee, Its Role And Reporting Structure {21a}
All data monitoring for this study is overseen by the project principal investigator (Hernández) and is independent from the study sponsor, nor are there any competing interests; however, there is no formal data committee. Instead, process evaluations will be conducted via regular research team meetings and a stakeholder advisory board consisting of non-research personnel.
Adverse Event Reporting And Harms {22}
All adverse events will be reported to the principal investigator and subsequently to the IRB overseeing the study.
Frequency And Plans For Auditing Trial Conduct {23}
Process evaluations includes checklists for intervention protocols and data collection packets, weekly meetings with data collectors, monthly investigator meetings and quarterly advisory group communications via formal meetings and/or online engagements such as 1–1 meetings and newsletter distributions. These activities will serve to periodically determine how well our study is operating. They also help to identify and resolve any problems or program needs. We will survey building ambassadors and document the number of door magnets that remain on display over time, while also tracking activities on virtual platforms. These activities serve to document the organizational and operational procedures. A monthly quality assurance check and report will be conducted by study staff to thoroughly review the following: recruitment, retention (follow-up), survey administration, quality of data collected, resource requirements and availability, barriers and facilitators to program implementation. We will note any deviations from the original design and changes in roles/responsibilities of partner and collaborators. The results improve our program’s sustainability, help identify lessons learned and best practices, and result in key replication recommendations.
Plans For Communicating Important Protocol Amendments To Relevant Parties (E.g. Trial Participants, Ethical Committees) {25}
All protocol modifications will be reported to the overseeing IRB, investigating team, additional research personnel, participants and registries in a timely fashion.
Dissemination plans {31a}
The study has a strong focus on dissemination both during and after completion of the study. Throughout the duration of the study, investigators and additional research personnel consistently collaborate with NYCHA and engage with members of a stakeholder advisory board. Board members include residents, resident leaders, NYCHA personnel, community-based organization leaders among other key stakeholders in the smoke-free realm. These engagements are part of our process evaluation efforts to ensure appropriate study implementation.
Additionally, post data collection and data analyses, manuscripts will be written for publication with study results.