- Email: [email protected]
The Equity Impact of Proactive Outreach to Smokers: Analysis of a Randomized Trial
RESEARCH ARTICLE
The Equity Impact of Proactive Outreach to Smokers: Analysis of a Randomized Trial 1 2XD1 XElisheva XD R. Danan, 4XD3 XMD, XD MPH,1,2 6XD5 XSteven XD S. Fu, 8XD7 XMD, XD MSCE,1,2 10XD9 XBarbara XD A. Clothier, 12XD XMS, XD 1,2 3 14XD3 XD XSiamak Noorbaloochi, 16XD5 XPhD, XD 18XD7 XPatrick XD J. Hammett, 20XD19 XMA, XD PhD,1,3 2XD1 XRachel XD Widome, 24XD3 XPhD, XD 1,2 26XD5 XDiana XD J. Burgess, 28XD7 XPhD XD
Introduction: Population-based smoking-cessation services tend to preferentially benefit high-SES smokers, potentially exacerbating disparities. Interventions that include proactive outreach, telephone counseling, and free or low-cost cessation medications may be more likely to help low-SES smokers quit. This analysis evaluated the role of SES in smokers’ response to a population-based proactive smoking-cessation intervention. Methods: This study, conducted in 2016 and 2017, was a secondary analysis of the Veterans Victory Over Tobacco Study, a multicenter pragmatic RCT of a proactive smoking-cessation intervention conducted from 2009 to 2011. Logistic regression modeling was used to test the effect of income or education level on 6-month prolonged abstinence at 1-year follow-up. Results: Of the 5,123 eligible, randomized participants, 2,565 (50%) reported their education level and 2,430 (47%) reported their income level. The interactions between education (p=0.07) or income (p=0.74) X treatment arm were not statistically significant at the 0.05 level. The largest effect sizes for the intervention were found among smokers in the lowest education category (11th grade), with a quit rate of 17.3% as compared with 5.7% in usual care (OR=3.5, 95% CI=1.4, 8.6) and in the lowest income range (<$10,000), with a quit rate of 18.7% as compared with 9.4% in usual care (OR=2.2, 95% CI=1.2, 4.0).
Conclusions: In a large, multicenter smoking-cessation trial, proactive outreach was associated with higher rates of prolonged abstinence among smokers at all SES levels. Proactive outreach interventions that integrate telephone-based care and facilitated cessation medication access have the potential to reduce socioeconomic disparities in quitting.
Trial registration: This study is registered at www.clinicaltrials.gov NCT00608426. Am J Prev Med 2018;55(4):506 516. Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine. This is an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)
INTRODUCTION
L
ow SES is a fundamental contributor to disease and premature death. As the prevalence of cigarette smoking declines in developed nations, SES-related smoking disparities expand.1,2 Low SES is associated with a higher prevalence of smoking,3 5 heavier use of cigarettes,6,7 and higher morbidity and mortality from smoking-related illness.8 Smoking cigarettes accounts for up to half the mortality difference between low- and high-SES men and women.9,10 506
From the 1Veterans Affairs Health Services Research & Development Center for Chronic Disease Outcomes Research, Minneapolis Veterans Affairs Health Care System, Minneapolis, Minnesota; 2Department of Medicine, University of Minnesota Medical School, Minneapolis, Minnesota; and 3 Division of Epidemiology, University of Minnesota School of Public Health, Minneapolis, Minnesota Address correspondence to: Elisheva R. Danan, MD, MPH, Center for Chronic Disease Outcomes Research, Minneapolis VA Health Care System, 1 Veterans Drive (152), Minneapolis MN 55417. E-mail: [email protected]. 0749-3797/$36.00 https://doi.org/10.1016/j.amepre.2018.05.023
Am J Prev Med 2018;55(4):506 516 Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine. This is an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)
507
Danan et al / Am J Prev Med 2018;55(4):506 516
Tobacco control efforts can inadvertently widen the socioeconomic gap by preferentially benefiting high-SES smokers. Phelan and Link’s Theory of Fundamental Causes explains that socioeconomic disparities in preventable diseases develop as a consequence of unequal access to knowledge or tools related to disease prevention.11,12 Traditional smoking-cessation services, including counseling and cessation medications, have historically been more accessible to high-SES smokers.7,13,14 Population-level interventions, such as massmedia campaigns or indoor-smoking bans, also often impact high-SES smokers preferentially.15 17 Only cigarette taxes have been shown to have a greater impact on tobacco use by low-SES smokers,18 though they achieve this by placing a heavier burden on the poor.19 An equity approach to tobacco control mandates that interventions aimed at shrinking the overall prevalence of smoking must simultaneously reduce smoking-related socioeconomic disparities.5 Yet for multiple reasons, promising interventions often have little impact on lowSES smokers. Though most studies find that low-SES smokers have similar interest in quitting and make the same number of quit attempts as high-SES smokers, they are far less likely to succeed.3,7,20,21 Individual-level barriers include heaviness of smoking, nicotine dependence, personal agency, and confidence to quit.21,22 Social and community resources, such as household factors, social support, and neighborhood disadvantage, also impact quit success.22 26 Low-SES smokers often begin smoking cessation treatment but are less likely to complete therapy.27 29 Certain features of tobacco control interventions may help low-SES smokers overcome obstacles. Phelan and Link30 advocate interventions that obviate the connection between having resources and accessing/using health promoting knowledge or treatment. For example, treatments should be affordable, easy to use, and readily accessible, overcoming advantages related to individual resources such as money, knowledge, and connections. Free or low-cost cessation medications and telephone counseling both improve access to evidence-based smoking-cessation therapies.5 Additionally, interventions should reach the entire population automatically, overcoming differential access to community-based resources. A proactive outreach approach to smoking cessation (as compared with the more common, reactive approach) involves contacting all smokers as a matter of course and offering help with quitting. The Veterans Victory Over Tobacco Study (Victory) was a pragmatic RCT of proactive versus usual care, in which the intervention combined proactive outreach with an offer of telephone or in-person smoking cessation counseling, as well as facilitated access to free or October 2018
31
low-cost cessation medications. The primary outcome paper reported an overall 2.6% population-level increase in 6-month prolonged abstinence at 1-year follow-up for smokers randomized to proactive care as compared with usual care.32 These population-level findings were similar to prior studies involving active recruitment of smokers to proactive telephone counseling interventions.33 However, it was unknown whether this intervention improved or widened SES-related smoking disparities. The current paper is a secondary analysis of the Victory Study with an equity focus. The primary question is whether the proactive care intervention had a differential effect on 6-month prolonged abstinence at 1-year follow-up for smokers at different SES levels. Secondary outcomes of interest include uptake of smoking-cessation treatments and quit attempts. The hypothesis being tested is that the proactive care intervention will help smokers at all SES levels, resulting in prolonged abstinence rates that do not vary by SES level.
METHODS Study Sample The Victory Study was a pragmatic RCT that received approval from all participating sites’ IRBs. Pragmatic trials use minimal inclusion/exclusion criteria to compare clinically relevant treatments under real-world conditions.34 Current smokers (aged 18 80 years) were identified using the U.S. Department of Veterans Affairs (VA) electronic medical record. Participants were recruited from October 2009 to September 2010 from four VA medical centers (New York City, New York; Jackson, Wyoming; Tampa, Florida; and Minneapolis, Minnesota) and follow-up was completed in November 2011. More details have been previously described.31,32 The current analysis was completed in 2017. The proactive care intervention combined an active recruitment strategy, proactive outreach (mailed materials followed by telephone outreach), with an offer of telephone smoking-cessation counseling or referral to in-person counseling. Telephone care included proactive counselor-initiated calls from counselors at the Minneapolis VA who were trained in motivational interviewing. Counselors also facilitated access to smoking-cessation pharmacotherapy through the participant’s VA provider. The usual-care group did not receive proactive outreach but did have access to smoking-cessation services through their local VA and state telephone quitline. VA administrative and healthcare utilization data were obtained from VA National Patient Care Databases. Survey data were collected at baseline and 1-year follow-up.
Measures SES was measured at baseline using self-reported education and income separately. Education levels included 11th grade, high school graduate or equivalent, some college, and college graduate or more. Income levels were defined by annual income <$10,000, $10,000 $20,000, $20,001 $40,000, $40,001 $60,000, and >$60,000. The primary outcome was self-reported 6-month prolonged abstinence at 1-year follow-up, and was assessed among all
508
Danan et al / Am J Prev Med 2018;55(4):506 516
participants, regardless of treatment utilization. Prolonged abstinence was defined as recommended for clinical trials by the Society for Research on Nicotine and Tobacco.35 Minor lapses in smoking over the 6 months prior to follow-up survey completion were permitted, but participants who reported smoking on 7 consecutive days or two consecutive weekends in the prior 6 months were not considered to have achieved prolonged abstinence. Self-report is the only method for obtaining prolonged abstinence and is the recommended method for measuring smoking status in large trials.36 It has been validated in a VA population.37 Secondary outcomes assessed included past-year uptake of smoking-cessation therapies and quit attempts measured at follow-up. Smoking cessation therapies examined were (1) selfreported receipt of phone counseling; (2) smoking-cessation medication dispensed by the VA, derived from VA administrative data (including bupropion, varenicline, or nicotine replacement therapy); and (3) combined therapy (use of behavioral counseling [telephone or in-person] and cessation medication [self-report or derived from VA administrative data]). These three cessation therapy outcomes were selected because they were previously shown to rise in response to proactive outreach in the Victory trial.32 The objective was to assess whether proactive outreach had a differential effect on uptake among participants at each SES level. Quit attempts were self-reported attempts to quit smoking intentionally for 24 hours.
Statistical Analysis The study sample was selected using a stratified random sampling plan (by site), and a completely randomized block design was used to randomly assign participants to the intervention or usual care. Accordingly, all estimations, testing, and modeling procedures are based on stratified analyses. A complete case analysis was performed and then augmented with non-ignorable modeling of nonresponse along with multiple imputation of the covariates, which were done to address possible biases due to missing outcomes or covariates. Logistic regression modeling was used to test the effect of SES (education or income level) on 6-month prolonged abstinence. Models included gender, age, treatment group, and site, along with the SES variable and the interaction of that variable X treatment group. Traditionally, in factorial designs, simple effects (i.e., within-group analyses) are only performed when a significant interaction term suggests that one factor has differential effects at different levels of another factor. However, unlike in a completely randomized factorial design, the grouping factors of interest in this secondary analysis (education and income) were not randomly assigned, creating the potential for confounding bias and imbalance because of observed and unobserved covariates. Therefore, proportion abstinent and treatment arm effect sizes (OR with 95% CI) were obtained within each SES level whether the interaction was significant or not, in order to further explore within-group relationships. To handle nonresponse for the primary outcome, 6-month prolonged abstinence, it was hypothesized that nonresponse might depend on the unobserved smoking status of the participant: that is, not missing at random (NMAR). The joint distribution of abstinence status and response status was modeled with a pair of logistic regression models and then an expectation maximization algorithm was used to find maximum
likelihood estimators, as described by Ibrahim and colleagues.38 This likelihood-based NMAR method creates two data sets, one that assumes all nonresponders are smokers and the other that assumes all nonresponders are quitters. Then, through a series of iterative weightings, it produces maximum likelihood estimates. Administrative data predictive of missingness was used to impute for missing education and income, producing five imputed data sets. The data analysis for this paper including the macro for likelihood-based NMAR modeling was generated using SAS/STAT software, version 9.2.
RESULTS Of the 5,123 eligible, randomized participants, the complete case samples for this secondary analysis consisted of those who provided follow-up smoking status along with either baseline education (n=2,565, 50%) or income (n=2,430, 47%). Within the five education and six income levels, response rates to the follow-up smoking status question ranged from 77.3% to 79.3% and 75.7% to 81.8%, respectively, and did not differ significantly by education or income level (p=0.63 and p=0.08). Education and income were highly correlated (p<0.001). Participant characteristics by SES level are presented in Tables 1 and 2. The interaction between education X treatment arm was not statistically significant (p=0.07) with respect to 6-month prolonged abstinence at the 0.05 level. Smokers at each education level had higher observed abstinence rates if randomized to proactive care versus usual care (Figure 1A). In individual subgroup analyses, the proactive care intervention had the largest effect size among smokers in the lowest education category (11th grade), with a quit rate of 17.3% as compared with 5.7% in usual care (OR=3.5, 95% CI=1.4, 8.6; Table 3). Analyses accounting for nonresponse using likelihood-based NMAR models and multiple imputation showed similar results to the main analysis, including finding the highest intervention effect size among the lowest education category. The interaction between income X treatment arm was not statistically significant (p=0.74). Smokers at each income level had higher observed abstinence rates if randomized to proactive care versus usual care (Figure 1B). The proactive care intervention had the largest effect size among smokers in the lowest income range (<$10,000), with a quit rate of 18.7% as compared with 9.4% in usual care (OR=2.2, 95% CI=1.2, 4.0; Table 3). Analyses accounting for nonresponse using likelihoodbased NMAR models and multiple imputation showed similar results to the main analysis, including finding the highest intervention effect size among the lowest income category. www.ajpmonline.org
509
Danan et al / Am J Prev Med 2018;55(4):506 516
Table 1. Sample Characteristics by Education Level
Characteristics
Total, n (%) or median (SD)
College, Some college, HS/GED, 11th grade, n (%) or median n (%) or median n (%) or median n (%) or median (SD) p-value (SD) (SD) (SD)
All participants 2,565 Treatment group Usual care 1,342 (52.32) Proactive care 1,223 (47.68) Demographic characteristics Age, years 60 (9.8) Race White 1,552 (60.51) Black 690 (26.9) Hispanic 163 (6.35) Other 160 (6.24) Male gender 2,431 (94.78) Married 1,260 (49.12) Smoking behaviors Cigarettes per day 10 909 (36.03) 11 20 1,070 (42.41) 21 544 (21.56) Time to first cigarette, minutes >31 710 (27.9) 6 30 1,324 (52.02) <5 511 (20.08) Social and environmental pressures Home smoking rules Not allowed 958 (39.39) anywhere Allowed some 528 (21.71) places/times Allowed anywhere 946 (38.9) Friends who smoke None 398 (16.46) Less than half 713 (29.49) About half 523 (21.63) More than half 463 (19.15) All 321 (13.28) People important to me want me to quit smoking Strongly disagree 496 (20.58) to neutral Somewhat agree 551 (22.86) Strongly agree 1,363 (56.56)
239
1,031
1,016
279
542 (52.57) 489 (47.43)
523 (51.48) 493 (48.52)
151 (54.12) 128 (45.88)
60 (9.3)
59 (10.2)
135 (56.49) 63 (26.36) 20 (8.37) 21 (8.79) 237 (99.16) 107 (44.77)
644 (62.46) 285 (27.64) 56 (5.43) 46 (4.46) 1,004 (97.38) 516 (50.05)
583 (57.38) 286 (28.15) 74 (7.28) 73 (7.19) 935 (92.03) 496 (48.82)
190 (68.1) 56 (20.07) 13 (4.66) 20 (7.17) 255 (91.4) 141 (50.54)
88 (37.93) 86 (37.07) 58 (25)
323 (31.92) 453 (44.76) 236 (23.32)
381 (37.95) 417 (41.53) 206 (20.52)
117 (42.55) 114 (41.45) 44 (16)
53 (22.27) 120 (50.42) 65 (27.31)
258 (25.32) 555 (54.47) 206 (20.22)
304 (30.07) 510 (50.45) 197 (19.49)
95 (34.3) 139 (50.18) 43 (15.52)
74 (33.48)
382 (39.38)
382 (39.02)
120 (45.8)
49 (22.17)
203 (20.93)
215 (21.96)
61 (23.28)
98 (44.34)
385 (39.69)
382 (39.02)
81 (30.92)
29 (13) 55 (24.66) 33 (14.8) 60 (26.91) 46 (20.63)
149 (15.47) 261 (27.1) 234 (24.3) 185 (19.21) 134 (13.91)
156 (16.1) 299 (30.86) 209 (21.57) 184 (18.99) 121 (12.49)
64 (24.33) 98 (37.26) 47 (17.87) 34 (12.93) 20 (7.6)
41 (18.64)
216 (22.48)
190 (19.59)
49 (18.92)
55 (25) 124 (56.36)
216 (22.48) 529 (55.05)
218 (22.47) 562 (57.94)
62 (23.94) 148 (57.14)
0.9 126 (52.72) 113 (47.28) 61 (7.7)
62 (11)
<0.001 0.002
<0.001 0.5 0.003
0.001
0.08
<0.001
0.6
Note: Boldface indicates statistical significance (p<0.05). HS/GED, high school/General Educational Development test.
Treatment uptake and quit attempts are reported in Appendix A and Appendix Tables 1 4. Briefly, the interactions between education or income X treatment arm with respect to all secondary outcomes related to treatment uptake and quit attempts were not statistically significant. Smokers at each level
October 2018
of education and income who were randomized to proactive care were more likely to report using telephone counseling or combined therapy than those randomized to usual care. VA smoking cessation medication use showed a weaker response to the proactive care intervention and quit attempts did
510
Table 2. Sample Characteristics by Income Level Characteristics
Total, n (%) or median (SD)
$10,000 $20,000, n (%) or median (SD)
$20,001 $40,000, n (%) or median (SD)
$40,001 $60,000, n (%) or median (SD)
>$60,000, n (%) or median (SD)
429
769
734
305
193
223 (51.98) 206 (48.02)
396 (51.5) 373 (48.5)
385 (52.45) 349 (47.55)
159 (52.13) 146 (47.87)
105 (54.4) 88 (45.6)
57 (9.1)
61 (9.3)
60 (10.2)
184 (42.89) 169 (39.39) 34 (7.93) 42 (9.79) 398 (92.77) 105 (24.48)
450 (58.52) 227 (29.52) 42 (5.46) 50 (6.5) 745 (96.88) 287 (37.32)
489 (66.62) 167 (22.75) 44 (5.99) 34 (4.63) 687 (93.6) 420 (57.22)
208 (68.2) 60 (19.67) 21 (6.89) 16 (5.25) 284 (93.11) 208 (68.2)
139 (72.02) 32 (16.58) 15 (7.77) 7 (3.63) 186 (96.37) 166 (86.01)
163 (38.81) 178 (42.38) 79 (18.81)
274 (36.15) 313 (41.29) 171 (22.56)
262 (36.34) 296 (41.05) 163 (22.61)
95 (31.46) 142 (47.02) 65 (21.52)
72 (37.5) 81 (42.19) 39 (20.31)
60 (10)
59 (10.9)
p-value — 0.97
<0.001 <0.001
0.005 <0.001 0.5
<0.001 104 (24.36) 210 (49.18) 113 (26.46)
198 (26.12) 391 (51.58) 169 (22.3)
209 (28.67) 376 (51.58) 144 (19.75)
90 (29.61) 169 (55.59) 45 (14.8)
73 (38.02) 98 (51.04) 21 (10.94)
124 (30.69)
247 (34.26)
278 (39.71)
141 (48.12)
114 (60.96)
89 (22.03)
145 (20.11)
160 (22.86)
63 (21.5)
38 (20.32)
191 (47.28)
329 (45.63)
262 (37.43)
89 (30.38)
35 (18.72)
<0.001
(continued on next page)
Danan et al / Am J Prev Med 2018;55(4):506 516
All participants 2,430 Treatment group Usual care 1,268 (52.18) Proactive care 1,162 (47.82) Demographic characteristics Age, years 60 (9.8) Race White 1,470 (60.49) Black 655 (26.95) Hispanic 156 (6.42) Other 149 (6.13) Male gender 2,300 (94.65) Married 1,186 (48.81) Smoking behaviors Cigarettes per day 10 866 (36.19) 11 20 1,010 (42.21) 21 517 (21.6) Time to first cigarette, minutes 31 674 (27.97) 6 30 1,244 (51.62) <5 492 (20.41) Social and environmental pressures Home smoking rules Not allowed 904 (39.22) anywhere Allowed some 495 (21.48) places/times Allowed anywhere 906 (39.31)
<$10,000, n (%) or median (SD)
www.ajpmonline.org
October 2018
0.3
<0.001
511
not differ significantly by treatment arm in most SES levels.
111 (60.33) 178 (61.81) 390 (54.47) 218 (53.69)
399 (57.74) Note: Boldface indicates statistical significance (p<0.05).
42 (22.83) 65 (22.57) 171 (23.88) 98 (24.14)
150 (21.71)
31 (16.85) 45 (15.63) 155 (21.65) 90 (22.17)
142 (20.55)
80 (11.13) 218 (30.32) 150 (20.86) 151 (21) 120 (16.69) 58 (14.39) 88 (21.84) 83 (20.6) 88 (21.84) 86 (21.34)
131 (18.79) 196 (28.12) 160 (22.96) 138 (19.8) 72 (10.33)
54 (18.82) 111 (38.68) 62 (21.6) 42 (14.63) 18 (6.27)
50 (27.03) 68 (36.76) 32 (17.3) 28 (15.14) 7 (3.78)
DISCUSSION
Friends who smoke None 373 (16.28) Less than half 681 (29.73) About half 487 (21.26) More than half 447 (19.51) All 303 (13.23) People important to me want me to quit smoking Strongly 463 (20.26) disagree to neutral Somewhat 526 (23.02) agree Strongly agree 1,296 (56.72)
<$10,000, n (%) or median (SD) Total, n (%) or median (SD) Characteristics
Table 2. Sample Characteristics by Income Level (continued)
$10,000 $20,000, n (%) or median (SD)
$20,001 $40,000, n (%) or median (SD)
$40,001 $60,000, n (%) or median (SD)
>$60,000, n (%) or median (SD)
p-value
Danan et al / Am J Prev Med 2018;55(4):506 516
In a large, multicenter smoking-cessation trial, a proactive outreach intervention was associated with improved rates of prolonged abstinence among smokers at all SES levels. The largest effect sizes for the intervention were among smokers at the lowest SES levels, who are historically the least likely to quit successfully. A proactive approach to smoking cessation may both lower the overall population-level smoking rate, as shown in the primary outcomes paper,32 and reduce SES-related smoking disparities, as shown here. These findings address a gap in evidence related to pro-equity smoking cessation support interventions. Cessation support interventions (including counseling, medications, or combined therapy) can have a net-negative impact on equity because they are less effective among low-SES smokers.18,39 A reactive treatment approach and the financial cost of treatment, often associated with lower use of cessation support services40 and lower quit rates in low-SES populations, were overcome in this study by (1) using proactive outreach and telephone-based treatment and (2) facilitating access to free or low-cost cessation medications. Traditional smoking-cessation interventions rely on motivated smokers to seek out services. Smokers who independently call quitlines for support41 or volunteer for intervention programs42 are more likely to come from a higher educational background than the general smoking population. By proactively calling smokers, the proactive outreach intervention addressed the fear of being judged, fear of failure, and lack of knowledge43 that may prevent low-SES smokers from seeking assistance. In addition, this telephone-based intervention did not rely on smokers to interface with the healthcare system directly, potentially increasing utilization by lowSES smokers.44 The cost of cessation medications is also a barrier to seeking and using care,45 and most lowincome U.S. smokers are unaware of insurance-subsidized treatment benefits.46 Integrating the offer of free or low-cost cessation medications with proactive outreach increased awareness of their availability, which is associated with increased utilization.46 Several studies have used proactive outreach to recruit smokers to quit,47 54 though none have evaluated the differential effect of proactive outreach interventions on smokers at different SES levels. Tzelepis et al.55 evaluated the overall relationship between SES and prolonged abstinence among actively recruited smokers. They found only shortterm cessation success for the primary intervention,54 and
512
Danan et al / Am J Prev Med 2018;55(4):506 516
Figure 1. Model-based proportion of smokers who reported 6-month prolonged abstinence at 1-year follow-up, by (A) education or (B) income level and treatment group. GED, General Educational Development test.
reported that among all smokers, being employed predicted one abstinence outcome at one timepoint, but other SES measures (education and private insurance) were not associated with abstinence.55 They did not assess the interaction between treatment group and SES. A pooled analysis of the intervention groups from five proactively recruited smoking-cessation studies (combined n=2,972) found a small but significant positive effect of education on point-prevalent abstinence at 12 and 24 months’ follow-up.56 Qualitatively, a similar pattern was observed among smokers of different education levels within the proactive care arm of this trial, for all except the lowest education level. The current study is enriched by the comparison with abstinence rates across education levels in the usual-care group arm. Proactive outreach has recently been shown to decrease smoking in low-SES populations in studies that
did not include higher-SES smokers.57,58 A comparable effect size at all SES levels was expected in the present study because knowledge, access, and utilization of the intervention were not contingent on individual or community resources. The nonsignificant interaction term supported this hypothesis, but it is surprising to find that the largest effect sizes for the intervention were in the lowest SES groups. Relatively high observed prolonged abstinence rates of 17% (11th grade education or less) or nearly 19% (less than $10,000 annual income) were unanticipated given that, consistent with historical patterns,6,7 low-SES smokers were heavier smokers with lower social and environmental pressures to quit. Variation in the utilization of cessation treatments or number of quit attempts (the secondary outcomes analyzed) did not explain this result. www.ajpmonline.org
513
Danan et al / Am J Prev Med 2018;55(4):506 516
Table 3. Six-Month Prolonged Abstinence (n, Model-Based %) by Treatment Arm and SES Level, With Model-Based OR Participants, n (%) SES level Education n 11th grade, n=239 HS/GED, n=1,031 Some college, n=1,016 College, n=279 Income, USD n <$10,000, n=429 $10,000 $20,000, n=769 $20,001 $40,000, n=734 $40,001 $60,000, n=305 >$60,000, n=193
Usual care
Proactive care
Complete casea model, OR (95% CI)
NMARb model, OR (95% CI)
1,342 6 (5.7) 47 (10.8) 36 (8.0) 12 (9.9)
1,223 16 (17.3) 46 (12.0) 60 (14.6) 23 (18.6)
2,565 3.46 (1.39, 8.59) 1.13 (0.76, 1.68) 1.96 (1.30, 2.96) 2.09 (1.03, 4.27)
5,123 2.05 (1.30, 3.24) 1.12 (0.91, 1.38) 1.42 (1.11, 1.82) 1.57 (0.87, 2.82)
1,268 18 (9.4) 29 (8.7) 31 (9.8) 9 (7.5) 12 (13.5)
1,162 32 (18.7) 36 (11.7) 45 (15.2) 11 (11.9) 12 (15.9)
2,430 2.23 (1.24, 3.98) 1.40 (0.86, 2.28) 1.65 (1.04, 2.62) 1.65 (0.74, 2.69) 1.21 (0.53, 2.78)
5,123 1.61 (1.09, 2.39) 1.17 (0.89, 1.54) 1.30 (1.00, 1.69) 1.54 (0.92, 2.56) 1.25 (0.70, 2.22)
Note: Boldface indicates statistical significance (p<0.05). a Complete case model: Usual care is reference group. Model controls for site, treatment arm, interaction between SES and treatment arm, gender, and age. Complete case includes follow-up survey responders for whom quit status is known. b NMAR model: Usual care is reference group. Likelihood-based not missing at random (NMAR) model accounts for non-response. HS/GED, high school/General Educational Development test.
Other possible explanations merit exploration. First, there may be pent-up demand for cessation assistance among low-SES smokers that leads to a preferential response to proactive outreach. Low-SES smokers making a first or second supported quit attempt may be more successful than high-SES smokers who have failed many previous quit attempts. One quitline evaluation found a stronger relationship between nicotine replacement therapy use and abstinence among low-SES smokers compared with high-SES smokers.43 On the other hand, higher-SES individuals who continue to smoke despite lower levels of nicotine dependence, high social and environmental pressures to quit, and many available quit resources may be resistant to additional outreach. Additionally, recent work by Hammett and colleagues59 confirmed that low-SES smokers often experience lower levels of perceived smoking-related stigma, or the social unacceptability of smoking. Low-SES smokers with a lower level of perceived smoking-related stigma at baseline reported higher abstinence rates at 1-year in response to a proactive outreach intervention than those with higher baseline levels of smoking-related stigma. It was hypothesized that proactive outreach contributed to denormalization of smoking among low-SES smokers whose social environments were conducive to tobacco use. In this trial, capturing smoking status in the electronic medical record made it possible to proactively reach out to all smokers, regardless of intent to quit, to facilitate cessation assistance. Although the VA healthcare system October 2018
was one of the first to adopt this practice, it is a recommended component of system-level change for smoking cessation that is becoming widespread.60 This strategy helped connect vulnerable subpopulations of smokers to evidence-based cessation treatments. Future trials of outreach to smokers, which are increasingly relying on mobile health technology,61,62 interactive voice response,57 and text messaging,61,63 65 should evaluate how those technologies address the needs of smokers at all SES levels.59
Limitations Prolonged abstinence is a self-reported measure that may be subject to social desirability bias. However, multiple studies have validated smoking self-report,66,67 including among VA smokers,37 and most research has failed to identify systematic differences in the risk of misclassification of smokers by SES level.68,69 Interaction tests may be underpowered, and the original trial was not designed to detect subgroup effects by SES level. The population of mostly male U.S. Veterans limits generalizability to female Veterans and non-Veterans. Finally, SES is a complex construct that is only partially captured by measured variables, such as education and income levels. Race, class, occupation, and subjective social standing all contribute to SES, but are outside the scope of this paper. For an analysis of the effect of proactive outreach on prolonged abstinence among African American and white smokers in the Victory trial, please see Burgess et al.70
514
Danan et al / Am J Prev Med 2018;55(4):506 516
CONCLUSIONS Future tobacco control policies and interventions must promote equity. Proactive outreach interventions have the potential to decrease population-level smoking prevalence while reducing smoking-related disparities.
ACKNOWLEDGMENTS The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the U.S. government. This study was funded by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, and Health Services Research and Development (IAB-05-303). ED collaborated on the conceptualization of this study, guided the analytic approach, interpreted the results, and wrote most sections of the draft manuscript. She oversaw all aspects of this study and takes full and final responsibility for the paper. SF secured funding for and designed the original Victory trial and collaborated on the conceptualization of the current analysis. SN and BC designed and carried out the statistical analysis and assisted with interpretation and wrote sections of the methods and results. RW and PH collaborated on the conceptualization of this study, critically revised the manuscript, and read and approved the final submitted version. DB is the senior author on the paper; she led the conceptualization and helped design the analytic approach, interpret the study findings, and revise the manuscript. All authors critically read, provided feedback, and edited the final paper. Registered in clinicaltrials.gov (NCT00608426). Earlier versions of this study were presented in a poster presentation at the Society for Research on Nicotine and Tobacco (Florence, Italy, March 2017); the Society for General Internal Medicine (Washington, DC, April 2017); and the Veterans Affairs Health Services Research and Development National Conference (Washington, DC, July 2017). No financial disclosures were reported by the authors of this paper.
SUPPLEMENTAL MATERIAL Supplemental materials associated with this article can be found in the online version at doi:10.1016/j. amepre.2018.05.023.
REFERENCES 1. Lopez AD, Collishaw NE, Piha T. A descriptive model of the cigarette epidemic in developed countries. Tob Control. 1994;3(3):242–247. https://doi.org/10.1136/tc.3.3.242. 2. Thun M, Peto R, Boreham J, Lopez AD. Stages of the cigarette epidemic on entering its second century. Tob Control. 2012;21(2):96–101. https://doi.org/10.1136/tobaccocontrol-2011-050294. 3. HHS. The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta, GA: HHS, CDC, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Published 2014.
4. Whyte B, Gordon D, Haw S, Fischbacher C, Scott N, Harrison R. An Atlas of Tobacco Smoking in Scotland: A Report Presenting Estimated Smoking Prevalence and Smoking Attributable Deaths Within Scotland. NHS Health Scotland. Published 2007. 5. Kunst A, Giskes K, Mackenbach J. Socio-economic Inequalities in Smoking in the European Union: Applying an Equity Lens to Tobacco Control Policies: for the EU Network on Interventions to Reduce Socioeconomic Inequalities in Health. Erasmus Medical Center Rotterdam, 2004. 6. Siahpush M, McNeill A, Borland R, Fong GT. Socioeconomic variations in nicotine dependence, self-efficacy, and intention to quit across four countries: findings from the International Tobacco Control (ITC) Four Country Survey. Tob Control. 2006;15(suppl 3):iii71–iii75. https://doi.org/10.1136/tc.2004.008763. 7. Kotz D, West R. Explaining the social gradient in smoking cessation: it’s not in the trying, but in the succeeding. Tob Control. 2009;18 (1):43–46. https://doi.org/10.1136/tc.2008.025981. 8. Mackenbach JP, Huisman M, Andersen O, et al. Inequalities in lung cancer mortality by the educational level in 10 European populations. Eur J Cancer. 2004;40(1):126–135. https://doi.org/10.1016/j. ejca.2003.10.018. 9. Jha P, Peto R, Zatonski W, Boreham J, Jarvis MJ, Lopez AD. Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet. 2006;368(9533):367–370. https:// doi.org/10.1016/S0140-6736(06)68975-7. 10. Gregoraci G, van Lenthe FJ, Artnik B, et al. Contribution of smoking to socioeconomic inequalities in mortality: a study of 14 European countries, 1990 2004. Tob Control. 2017;26(3):260–268. https://doi. org/10.1136/tobaccocontrol-2015-052766. 11. Link BG, Phelan J. Social conditions as fundamental causes of disease. J Health Soc Behav. 1995:80 94. https://doi.org/10.2307/2626958. 12. Phelan JC, Link BG. Controlling disease and creating disparities: a fundamental cause perspective. J Gerontol B Psychol Sci Soc Sci. 2005;60(Special Issue 2):S27–S33. 13. Bauld L, Judge K, Platt S. Assessing the impact of smoking cessation services on reducing health inequalities in England: observational study. Tob Control. 2007;16(6):400–404. https://doi.org/10.1136/ tc.2007.021626. 14. Browning KK, Ferketich AK, Salsberry PJ, Wewers ME. Socioeconomic disparity in provider-delivered assistance to quit smoking. Nicotine Tob Res. 2008;10(1):55–61. https://doi.org/10.1080/14622200701704905. 15. Niederdeppe J, Fiore MC, Baker TB, Smith SS. Smoking-cessation media campaigns and their effectiveness among socioeconomically advantaged and disadvantaged populations. Am J Public Health. 2008;98(5):916–924. https://doi.org/10.2105/AJPH.2007.117499. 16. Jarvis MJ, Wardle J. Social Patterning of Individual Health Behaviours: The Case of Cigarette Smoking. Oxford, UK: Oxford University Press; 2005. https://doi.org/10.1093/acprof:oso/9780198565895.003.11. 17. Thomas S, Fayter D, Misso K, et al. Population tobacco control interventions and their effects on social inequalities in smoking: systematic review. Tob Control. 2008;17(4):230–237. https://doi.org/10.1136/ tc.2007.023911. 18. Hill S, Amos A, Clifford D, Platt S. Impact of tobacco control interventions on socioeconomic inequalities in smoking: review of the evidence. Tob Control. 2014;23(e2):e89–e97. https://doi.org/10.1136/ tobaccocontrol-2013-051110. 19. Remler DK. Poor smokers, poor quitters, and cigarette tax regressivity. Am J Public Health. 2004;94(2):225–229. https://doi.org/10.2105/ AJPH.94.2.225. 20. West R, McEwen A, Bolling K, Owen L. Smoking cessation and smoking patterns in the general population: a 1-year follow-up. Addiction. 2001;96 (6):891–902. https://doi.org/10.1046/j.1360-0443.2001.96689110.x. 21. Vangeli E, Stapleton J, Smit ES, Borland R, West R. Predictors of attempts to stop smoking and their success in adult general
www.ajpmonline.org
Danan et al / Am J Prev Med 2018;55(4):506 516
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
population samples: a systematic review. Addiction. 2011;106 (12):2110–2121. https://doi.org/10.1111/j.1360-0443.2011.03565.x. Chandola T, Head J, Bartley M. Socio-demographic predictors of quitting smoking: how important are household factors. Addiction. 2004;99 (6):770–777. https://doi.org/10.1111/j.1360-0443.2004.00756.x. Giskes K, van Lenthe FJ, Turrell G, Brug J, Mackenbach JP. Smokers living in deprived areas are less likely to quit: a longitudinal follow-up. Tob Control. 2006;15(6):485–488. https://doi.org/10.1136/tc.2006.015750. Lacey LP, Manfredi C, Balch G, Warnecke RB, Allen K, Edwards C. Social support in smoking cessation among black women in Chicago public housing. Public Health Rep. 1993;108(3):387–394. Mermelstein R, Cohen S, Lichtenstein E, Baer JS, Kamarck T. Social support and smoking cessation and maintenance. J Consult Clin Psychol. 1986;54(4):447–453. https://doi.org/10.1037/0022006X.54.4.447. Businelle MS, Kendzor DE, Reitzel LR, et al. Mechanisms linking socioeconomic status to smoking cessation: a structural equation modeling approach. Health Psychol. 2010;29(3):262–273. https://doi. org/10.1037/a0019285. Lam T-H, Abdullah ASM, Chan SS, Hedley AJ. Adherence to nicotine replacement therapy versus quitting smoking among Chinese smokers: a preliminary investigation. Psychopharmacology. 2005;177 (4):400–408. https://doi.org/10.1007/s00213-004-1971-y. Nevid JS, Javier RA, Moulton III JL. Factors predicting participant attrition in a community-based, culturally specific smokingcessation program for Hispanic smokers. Health Psychol. 1996;15 (3):226–229. https://doi.org/10.1037/0278-6133.15.3.226. Hiscock R, Judge K, Bauld L. Social inequalities in quitting smoking: what factors mediate the relationship between socioeconomic position and smoking cessation?. J Public Health (Oxf). 2011;33(1):39–47. https://doi.org/10.1093/pubmed/fdq097. Phelan JC, Link BG, Tehranifar P. Social conditions as fundamental causes of health inequalities: theory, evidence, and policy implications. J Health Soc Behav. 2010;51(1):S28–S40. (suppl) https://doi.org/ 10.1177/0022146510383498. Fu SS, van Ryn M, Sherman SE, et al. Population-based tobacco treatment: study design of a randomized controlled trial. BMC Public Health. 2012;12:159 https://doi.org/10.1186/1471-2458-12-159. Fu SS, van Ryn M, Sherman SE, et al. Proactive tobacco treatment and population-level cessation: a pragmatic randomized clinical trial. JAMA Intern Med. 2014;174(5):671–677. https://doi.org/10.1001/ jamainternmed.2014.177. Tzelepis F, Paul CL, Walsh RA, McElduff P, Knight J. Proactive telephone counseling for smoking cessation: meta-analyses by recruitment channel and methodological quality. J Natl Cancer Inst. 2011;103(12):922–941. https://doi.org/10.1093/jnci/djr169. Zwarenstein M, Treweek S, Gagnier JJ, et al. Improving the reporting of pragmatic trials: an extension of the CONSORT statement. BMJ. 2008;337:a2390 https://doi.org/10.1136/bmj.a2390. Hughes JR, Keely JP, Niaura RS, Ossip-Klein DJ, Richmond RL, Swan GE. Measures of abstinence in clinical trials: issues and recommendations. Nicotine Tob Res. 2003;5(1):13–25. https://doi.org/10.1080/ 1462220031000070552. Benowitz NL, Jacob III P, Ahijevych K, et al. Biochemical verification of tobacco use and cessation. Nicotine Tob Res. 2002;4(2):149–159. https://doi.org/10.1080/14622200210123581. Noonan D, Jiang Y, Duffy SA. Utility of biochemical verification of tobacco cessation in the Department of Veterans Affairs. Addict Behav. 2013;38(3):1792–1795. https://doi.org/10.1016/j.addbeh.2012.11.006. Ibrahim JG, Chen MH, Lipsitz SR. Missing responses in generalised linear mixed models when the missing data mechanism is nonignorable. Biometrika. 2001;88(2):551–564. https://doi.org/10.1093/biomet/88.2.551. Main C, Thomas S, Ogilvie D, et al. Population tobacco control interventions and their effects on social inequalities in smoking: placing an
October 2018
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
515
equity lens on existing systematic reviews. BMC Public Health. 2008;8:178. https://doi.org/10.1186/1471-2458-8-178. Roddy E, Antoniak M, Britton J, Molyneux A, Lewis S. Barriers and motivators to gaining access to smoking cessation services amongst deprived smokers—a qualitative study. BMC Health Serv Res. 2006;6:147 https://doi.org/10.1186/1472-6963-6-147. Tzelepis F, Paul CL, Walsh RA, et al. Telephone recruitment into a randomized controlled trial of quitline support. Am J Prev Med. 2009;37(4):324–329. https://doi.org/10.1016/j.amepre.2009.05.022. Monso E, Campbell J, Tønnesen P, Gustavsson G, Morera J. Sociodemographic predictors of success in smoking intervention. Tob Control. 2001;10(2):165–169. https://doi.org/10.1136/tc.10.2.165. Murray RL, Bauld L, Hackshaw LE, McNeill A. Improving access to smoking cessation services for disadvantaged groups: a systematic review. J Public Health. 2009;31(2):258–277. https://doi.org/10.1093/ pubmed/fdp008. Bryant J, Bonevski B, Paul C, McElduff P, Attia J. A systematic review and meta-analysis of the effectiveness of behavioural smoking cessation interventions in selected disadvantaged groups. Addiction. 2011;106 (9):1568–1585. https://doi.org/10.1111/j.1360-0443.2011.03467.x. Beauchamp A, Peeters A, Tonkin A, Turrell G. Best practice for prevention and treatment of cardiovascular disease through an equity lens: a review. Eur J Cardiovasc Prev Rehabil. 2010;17(5):599–606. https://doi.org/10.1097/HJR.0b013e328339cc99. McMenamin SB, Halpin HA, Bellows NM. Knowledge of Medicaid coverage and effectiveness of smoking treatments. Am J Prev Med. 2006;31(5):369–374. https://doi.org/10.1016/j.amepre.2006.07.015. Abdullah ASM, Mak YW, Loke AY, Lam TH. Smoking cessation intervention in parents of young children: a randomised controlled trial. Addiction. 2005;100(11):1731–1740. https://doi.org/10.1111/ j.1360-0443.2005.01231.x. Aveyard P, Griffin C, Lawrence T, Cheng K. A controlled trial of an expert system and self-help manual intervention based on the stages of change versus standard self-help materials in smoking cessation. Addiction. 2003;98(3):345 https://doi.org/10.1046/j.1360-0443.2003.00302.x. McBride CM, Scholes D, Grothaus LC, Curry SJ, Ludman E, Albright J. Evaluation of a minimal self-help smoking cessation intervention following cervical cancer screening. Prev Med. 1999;29(2):133–138. https://doi.org/10.1006/pmed.1999.0514. Curry SJ, McBride C, Grothaus LC, Louie D, Wagner EH. A randomized trial of self-help materials, personalized feedback, and telephone counseling with nonvolunteer smokers. J Consult Clin Psychol. 1995;63(6):1005 https://doi.org/10.1037/0022-006X.63.6.1005. Lichtenstein E, Boles SM, Lee ME, Hampson SE, Glasgow RE, Fellows J. Using radon risk to motivate smoking reduction II: randomized evaluation of brief telephone counseling and a targeted video. Health Educ Res. 2007;23(2):191–201. https://doi.org/10.1093/her/cym016. Lichtenstein E, Andrews JA, Lee ME, Glasgow RE, Hampson SE. Using radon risk to motivate smoking reduction: evaluation of written materials and brief telephone counselling. Tob Control. 2000;9(3):320– 326. https://doi.org/10.1136/tc.9.3.320. Prochaska JO, Velicer WF, Fava JL, et al. Counselor and stimulus control enhancements of a stage-matched expert system intervention for smokers in a managed care setting. Prev Med. 2001;32(1):23–32. https://doi.org/10.1006/pmed.2000.0767. Tzelepis F, Paul CL, Wiggers J, et al. A randomised controlled trial of proactive telephone counselling on cold-called smokers’ cessation rates. Tob Control. 2011;20(1):40–46. https://doi.org/10.1136/tc.2010.035956. Tzelepis F, Paul CL, Walsh RA, Wiggers J, Duncan SL, Knight J. Predictors of abstinence among smokers recruited actively to quitline support. Addiction. 2013;108(1):181–185. https://doi.org/10.1111/j.13600443.2012.03998.x. Velicer WF, Redding CA, Sun X, Prochaska JO. Demographic variables, smoking variables, and outcome across five studies. Health Psychol. 2007;26(3):278 https://doi.org/10.1037/0278-6133.26.3.278.
516
Danan et al / Am J Prev Med 2018;55(4):506 516
57. Haas JS, Linder JA, Park ER, et al. Proactive tobacco cessation outreach to smokers of low socioeconomic status: a randomized clinical trial. JAMA Intern Med. 2015;175(2):218–226. https://doi.org/10.1001/ jamainternmed.2014.6674. 58. Fu SS, van Ryn M, Nelson D, et al. Proactive tobacco treatment offering free nicotine replacement therapy and telephone counselling for socioeconomically disadvantaged smokers: a randomised clinical trial. Thorax. 2016;71(5):446–453. https://doi.org/10.1136/thoraxjnl-2015-207904. 59. Hammett P, Fu SS, Nelson D, et al. A proactive smoking cessation intervention for socioeconomically disadvantaged smokers: the role of smoking-related stigma. Nicotine Tob Res. 2018;20(3):286–294. https://doi.org/10.1093/ntr/ntx085. 60. Thomas D, Abramson MJ, Bonevski B, George J. System change interventions for smoking cessation. Cochrane Database Syst Rev. 2017;2: CD010742 https://doi.org/10.1002/14651858.CD010742.pub2. 61. Free C, Phillips G, Galli L, et al. The effectiveness of mobile-health technology-based health behaviour change or disease management interventions for health care consumers: a systematic review. PLoS Med. 2013;10 (1):e1001362 https://doi.org/10.1371/journal.pmed.1001362. 62. Heffner JL, Mull KE. Smartphone ownership among U.S. adult cigarette smokers: 2014 Health Information National Trends Survey (HINTS) data. J Med Internet Res. 2017;19(8):e305 https://doi.org/10.2196/jmir.7953. 63. Whittaker R, Borland R, Bullen C, Lin RB, McRobbie H, Rodgers A. Mobile phone-based interventions for smoking cessation. Cochrane Database Syst Rev. 2009;4:CD006611 https://doi.org/10.1002/14651858.CD006611.pub2.
64. Armanasco AA, Miller YD, Fjeldsoe BS, Marshall AL. Preventive health behavior change text message interventions: a meta-analysis. Am J Prev Med. 2017;52(3):391–402. https://doi.org/10.1016/j. amepre.2016.10.042. 65. Abroms LC, Johnson PR, Leavitt LE, et al. A randomized trial of text messaging for smoking cessation in pregnant women. Am J Prev Med. 2017;53(6):781–790. https://doi.org/10.1016/j.amepre.2017.08.002. 66. Yeager DS, Krosnick JA. The validity of self-reported nicotine product use in the 2001 2008 National Health and Nutrition Examination Survey. Med Care. 2010;48(12):1128–1132. https://doi.org/10.1097/ MLR.0b013e3181ef9948. 67. Patrick DL, Cheadle A, Thompson DC, Diehr P, Koepsell T, Kinne S. The validity of self-reported smoking: a review and meta-analysis. Am J Public Health. 1994;84(7):1086–1093. https://doi.org/10.2105/AJPH.84.7.1086. 68. Vartiainen E, Sepp€al€a T, Lillsunde P, Puska P. Validation of self reported smoking by serum cotinine measurement in a communitybased study. J Epidemiol Community Health. 2002;56(3):167–170. https://doi.org/10.1136/jech.56.3.167. 69. Graham H, Owen L. Are there socioeconomic differentials in underreporting of smoking in pregnancy. Tob Control. 2003;12(4):434 https://doi.org/10.1136/tc.12.4.434. 70. Burgess DJ, Van Ryn M, Noorbaloochi S, et al. Smoking cessation among African American and white smokers in the Veterans Affairs health care system. Am J Public Health. 2014;104(suppl 4):S580–S587. https://doi.org/10.2105/AJPH.2014.302023.
www.ajpmonline.org
The Equity Impact of Proactive Outreach to Smokers: Analysis of a Randomized Trial 1 2XD1 XElisheva XD R. Danan, 4XD3 XMD, XD MPH,1,2 6XD5 XSteven XD S. Fu, 8XD7 XMD, XD MSCE,1,2 10XD9 XBarbara XD A. Clothier, 12XD XMS, XD 1,2 3 14XD3 XD XSiamak Noorbaloochi, 16XD5 XPhD, XD 18XD7 XPatrick XD J. Hammett, 20XD19 XMA, XD PhD,1,3 2XD1 XRachel XD Widome, 24XD3 XPhD, XD 1,2 26XD5 XDiana XD J. Burgess, 28XD7 XPhD XD
Introduction: Population-based smoking-cessation services tend to preferentially benefit high-SES smokers, potentially exacerbating disparities. Interventions that include proactive outreach, telephone counseling, and free or low-cost cessation medications may be more likely to help low-SES smokers quit. This analysis evaluated the role of SES in smokers’ response to a population-based proactive smoking-cessation intervention. Methods: This study, conducted in 2016 and 2017, was a secondary analysis of the Veterans Victory Over Tobacco Study, a multicenter pragmatic RCT of a proactive smoking-cessation intervention conducted from 2009 to 2011. Logistic regression modeling was used to test the effect of income or education level on 6-month prolonged abstinence at 1-year follow-up. Results: Of the 5,123 eligible, randomized participants, 2,565 (50%) reported their education level and 2,430 (47%) reported their income level. The interactions between education (p=0.07) or income (p=0.74) X treatment arm were not statistically significant at the 0.05 level. The largest effect sizes for the intervention were found among smokers in the lowest education category (11th grade), with a quit rate of 17.3% as compared with 5.7% in usual care (OR=3.5, 95% CI=1.4, 8.6) and in the lowest income range (<$10,000), with a quit rate of 18.7% as compared with 9.4% in usual care (OR=2.2, 95% CI=1.2, 4.0).
Conclusions: In a large, multicenter smoking-cessation trial, proactive outreach was associated with higher rates of prolonged abstinence among smokers at all SES levels. Proactive outreach interventions that integrate telephone-based care and facilitated cessation medication access have the potential to reduce socioeconomic disparities in quitting.
Trial registration: This study is registered at www.clinicaltrials.gov NCT00608426. Am J Prev Med 2018;55(4):506 516. Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine. This is an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)
INTRODUCTION
L
ow SES is a fundamental contributor to disease and premature death. As the prevalence of cigarette smoking declines in developed nations, SES-related smoking disparities expand.1,2 Low SES is associated with a higher prevalence of smoking,3 5 heavier use of cigarettes,6,7 and higher morbidity and mortality from smoking-related illness.8 Smoking cigarettes accounts for up to half the mortality difference between low- and high-SES men and women.9,10 506
From the 1Veterans Affairs Health Services Research & Development Center for Chronic Disease Outcomes Research, Minneapolis Veterans Affairs Health Care System, Minneapolis, Minnesota; 2Department of Medicine, University of Minnesota Medical School, Minneapolis, Minnesota; and 3 Division of Epidemiology, University of Minnesota School of Public Health, Minneapolis, Minnesota Address correspondence to: Elisheva R. Danan, MD, MPH, Center for Chronic Disease Outcomes Research, Minneapolis VA Health Care System, 1 Veterans Drive (152), Minneapolis MN 55417. E-mail: [email protected]. 0749-3797/$36.00 https://doi.org/10.1016/j.amepre.2018.05.023
Am J Prev Med 2018;55(4):506 516 Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine. This is an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)
507
Danan et al / Am J Prev Med 2018;55(4):506 516
Tobacco control efforts can inadvertently widen the socioeconomic gap by preferentially benefiting high-SES smokers. Phelan and Link’s Theory of Fundamental Causes explains that socioeconomic disparities in preventable diseases develop as a consequence of unequal access to knowledge or tools related to disease prevention.11,12 Traditional smoking-cessation services, including counseling and cessation medications, have historically been more accessible to high-SES smokers.7,13,14 Population-level interventions, such as massmedia campaigns or indoor-smoking bans, also often impact high-SES smokers preferentially.15 17 Only cigarette taxes have been shown to have a greater impact on tobacco use by low-SES smokers,18 though they achieve this by placing a heavier burden on the poor.19 An equity approach to tobacco control mandates that interventions aimed at shrinking the overall prevalence of smoking must simultaneously reduce smoking-related socioeconomic disparities.5 Yet for multiple reasons, promising interventions often have little impact on lowSES smokers. Though most studies find that low-SES smokers have similar interest in quitting and make the same number of quit attempts as high-SES smokers, they are far less likely to succeed.3,7,20,21 Individual-level barriers include heaviness of smoking, nicotine dependence, personal agency, and confidence to quit.21,22 Social and community resources, such as household factors, social support, and neighborhood disadvantage, also impact quit success.22 26 Low-SES smokers often begin smoking cessation treatment but are less likely to complete therapy.27 29 Certain features of tobacco control interventions may help low-SES smokers overcome obstacles. Phelan and Link30 advocate interventions that obviate the connection between having resources and accessing/using health promoting knowledge or treatment. For example, treatments should be affordable, easy to use, and readily accessible, overcoming advantages related to individual resources such as money, knowledge, and connections. Free or low-cost cessation medications and telephone counseling both improve access to evidence-based smoking-cessation therapies.5 Additionally, interventions should reach the entire population automatically, overcoming differential access to community-based resources. A proactive outreach approach to smoking cessation (as compared with the more common, reactive approach) involves contacting all smokers as a matter of course and offering help with quitting. The Veterans Victory Over Tobacco Study (Victory) was a pragmatic RCT of proactive versus usual care, in which the intervention combined proactive outreach with an offer of telephone or in-person smoking cessation counseling, as well as facilitated access to free or October 2018
31
low-cost cessation medications. The primary outcome paper reported an overall 2.6% population-level increase in 6-month prolonged abstinence at 1-year follow-up for smokers randomized to proactive care as compared with usual care.32 These population-level findings were similar to prior studies involving active recruitment of smokers to proactive telephone counseling interventions.33 However, it was unknown whether this intervention improved or widened SES-related smoking disparities. The current paper is a secondary analysis of the Victory Study with an equity focus. The primary question is whether the proactive care intervention had a differential effect on 6-month prolonged abstinence at 1-year follow-up for smokers at different SES levels. Secondary outcomes of interest include uptake of smoking-cessation treatments and quit attempts. The hypothesis being tested is that the proactive care intervention will help smokers at all SES levels, resulting in prolonged abstinence rates that do not vary by SES level.
METHODS Study Sample The Victory Study was a pragmatic RCT that received approval from all participating sites’ IRBs. Pragmatic trials use minimal inclusion/exclusion criteria to compare clinically relevant treatments under real-world conditions.34 Current smokers (aged 18 80 years) were identified using the U.S. Department of Veterans Affairs (VA) electronic medical record. Participants were recruited from October 2009 to September 2010 from four VA medical centers (New York City, New York; Jackson, Wyoming; Tampa, Florida; and Minneapolis, Minnesota) and follow-up was completed in November 2011. More details have been previously described.31,32 The current analysis was completed in 2017. The proactive care intervention combined an active recruitment strategy, proactive outreach (mailed materials followed by telephone outreach), with an offer of telephone smoking-cessation counseling or referral to in-person counseling. Telephone care included proactive counselor-initiated calls from counselors at the Minneapolis VA who were trained in motivational interviewing. Counselors also facilitated access to smoking-cessation pharmacotherapy through the participant’s VA provider. The usual-care group did not receive proactive outreach but did have access to smoking-cessation services through their local VA and state telephone quitline. VA administrative and healthcare utilization data were obtained from VA National Patient Care Databases. Survey data were collected at baseline and 1-year follow-up.
Measures SES was measured at baseline using self-reported education and income separately. Education levels included 11th grade, high school graduate or equivalent, some college, and college graduate or more. Income levels were defined by annual income <$10,000, $10,000 $20,000, $20,001 $40,000, $40,001 $60,000, and >$60,000. The primary outcome was self-reported 6-month prolonged abstinence at 1-year follow-up, and was assessed among all
508
Danan et al / Am J Prev Med 2018;55(4):506 516
participants, regardless of treatment utilization. Prolonged abstinence was defined as recommended for clinical trials by the Society for Research on Nicotine and Tobacco.35 Minor lapses in smoking over the 6 months prior to follow-up survey completion were permitted, but participants who reported smoking on 7 consecutive days or two consecutive weekends in the prior 6 months were not considered to have achieved prolonged abstinence. Self-report is the only method for obtaining prolonged abstinence and is the recommended method for measuring smoking status in large trials.36 It has been validated in a VA population.37 Secondary outcomes assessed included past-year uptake of smoking-cessation therapies and quit attempts measured at follow-up. Smoking cessation therapies examined were (1) selfreported receipt of phone counseling; (2) smoking-cessation medication dispensed by the VA, derived from VA administrative data (including bupropion, varenicline, or nicotine replacement therapy); and (3) combined therapy (use of behavioral counseling [telephone or in-person] and cessation medication [self-report or derived from VA administrative data]). These three cessation therapy outcomes were selected because they were previously shown to rise in response to proactive outreach in the Victory trial.32 The objective was to assess whether proactive outreach had a differential effect on uptake among participants at each SES level. Quit attempts were self-reported attempts to quit smoking intentionally for 24 hours.
Statistical Analysis The study sample was selected using a stratified random sampling plan (by site), and a completely randomized block design was used to randomly assign participants to the intervention or usual care. Accordingly, all estimations, testing, and modeling procedures are based on stratified analyses. A complete case analysis was performed and then augmented with non-ignorable modeling of nonresponse along with multiple imputation of the covariates, which were done to address possible biases due to missing outcomes or covariates. Logistic regression modeling was used to test the effect of SES (education or income level) on 6-month prolonged abstinence. Models included gender, age, treatment group, and site, along with the SES variable and the interaction of that variable X treatment group. Traditionally, in factorial designs, simple effects (i.e., within-group analyses) are only performed when a significant interaction term suggests that one factor has differential effects at different levels of another factor. However, unlike in a completely randomized factorial design, the grouping factors of interest in this secondary analysis (education and income) were not randomly assigned, creating the potential for confounding bias and imbalance because of observed and unobserved covariates. Therefore, proportion abstinent and treatment arm effect sizes (OR with 95% CI) were obtained within each SES level whether the interaction was significant or not, in order to further explore within-group relationships. To handle nonresponse for the primary outcome, 6-month prolonged abstinence, it was hypothesized that nonresponse might depend on the unobserved smoking status of the participant: that is, not missing at random (NMAR). The joint distribution of abstinence status and response status was modeled with a pair of logistic regression models and then an expectation maximization algorithm was used to find maximum
likelihood estimators, as described by Ibrahim and colleagues.38 This likelihood-based NMAR method creates two data sets, one that assumes all nonresponders are smokers and the other that assumes all nonresponders are quitters. Then, through a series of iterative weightings, it produces maximum likelihood estimates. Administrative data predictive of missingness was used to impute for missing education and income, producing five imputed data sets. The data analysis for this paper including the macro for likelihood-based NMAR modeling was generated using SAS/STAT software, version 9.2.
RESULTS Of the 5,123 eligible, randomized participants, the complete case samples for this secondary analysis consisted of those who provided follow-up smoking status along with either baseline education (n=2,565, 50%) or income (n=2,430, 47%). Within the five education and six income levels, response rates to the follow-up smoking status question ranged from 77.3% to 79.3% and 75.7% to 81.8%, respectively, and did not differ significantly by education or income level (p=0.63 and p=0.08). Education and income were highly correlated (p<0.001). Participant characteristics by SES level are presented in Tables 1 and 2. The interaction between education X treatment arm was not statistically significant (p=0.07) with respect to 6-month prolonged abstinence at the 0.05 level. Smokers at each education level had higher observed abstinence rates if randomized to proactive care versus usual care (Figure 1A). In individual subgroup analyses, the proactive care intervention had the largest effect size among smokers in the lowest education category (11th grade), with a quit rate of 17.3% as compared with 5.7% in usual care (OR=3.5, 95% CI=1.4, 8.6; Table 3). Analyses accounting for nonresponse using likelihood-based NMAR models and multiple imputation showed similar results to the main analysis, including finding the highest intervention effect size among the lowest education category. The interaction between income X treatment arm was not statistically significant (p=0.74). Smokers at each income level had higher observed abstinence rates if randomized to proactive care versus usual care (Figure 1B). The proactive care intervention had the largest effect size among smokers in the lowest income range (<$10,000), with a quit rate of 18.7% as compared with 9.4% in usual care (OR=2.2, 95% CI=1.2, 4.0; Table 3). Analyses accounting for nonresponse using likelihoodbased NMAR models and multiple imputation showed similar results to the main analysis, including finding the highest intervention effect size among the lowest income category. www.ajpmonline.org
509
Danan et al / Am J Prev Med 2018;55(4):506 516
Table 1. Sample Characteristics by Education Level
Characteristics
Total, n (%) or median (SD)
College, Some college, HS/GED, 11th grade, n (%) or median n (%) or median n (%) or median n (%) or median (SD) p-value (SD) (SD) (SD)
All participants 2,565 Treatment group Usual care 1,342 (52.32) Proactive care 1,223 (47.68) Demographic characteristics Age, years 60 (9.8) Race White 1,552 (60.51) Black 690 (26.9) Hispanic 163 (6.35) Other 160 (6.24) Male gender 2,431 (94.78) Married 1,260 (49.12) Smoking behaviors Cigarettes per day 10 909 (36.03) 11 20 1,070 (42.41) 21 544 (21.56) Time to first cigarette, minutes >31 710 (27.9) 6 30 1,324 (52.02) <5 511 (20.08) Social and environmental pressures Home smoking rules Not allowed 958 (39.39) anywhere Allowed some 528 (21.71) places/times Allowed anywhere 946 (38.9) Friends who smoke None 398 (16.46) Less than half 713 (29.49) About half 523 (21.63) More than half 463 (19.15) All 321 (13.28) People important to me want me to quit smoking Strongly disagree 496 (20.58) to neutral Somewhat agree 551 (22.86) Strongly agree 1,363 (56.56)
239
1,031
1,016
279
542 (52.57) 489 (47.43)
523 (51.48) 493 (48.52)
151 (54.12) 128 (45.88)
60 (9.3)
59 (10.2)
135 (56.49) 63 (26.36) 20 (8.37) 21 (8.79) 237 (99.16) 107 (44.77)
644 (62.46) 285 (27.64) 56 (5.43) 46 (4.46) 1,004 (97.38) 516 (50.05)
583 (57.38) 286 (28.15) 74 (7.28) 73 (7.19) 935 (92.03) 496 (48.82)
190 (68.1) 56 (20.07) 13 (4.66) 20 (7.17) 255 (91.4) 141 (50.54)
88 (37.93) 86 (37.07) 58 (25)
323 (31.92) 453 (44.76) 236 (23.32)
381 (37.95) 417 (41.53) 206 (20.52)
117 (42.55) 114 (41.45) 44 (16)
53 (22.27) 120 (50.42) 65 (27.31)
258 (25.32) 555 (54.47) 206 (20.22)
304 (30.07) 510 (50.45) 197 (19.49)
95 (34.3) 139 (50.18) 43 (15.52)
74 (33.48)
382 (39.38)
382 (39.02)
120 (45.8)
49 (22.17)
203 (20.93)
215 (21.96)
61 (23.28)
98 (44.34)
385 (39.69)
382 (39.02)
81 (30.92)
29 (13) 55 (24.66) 33 (14.8) 60 (26.91) 46 (20.63)
149 (15.47) 261 (27.1) 234 (24.3) 185 (19.21) 134 (13.91)
156 (16.1) 299 (30.86) 209 (21.57) 184 (18.99) 121 (12.49)
64 (24.33) 98 (37.26) 47 (17.87) 34 (12.93) 20 (7.6)
41 (18.64)
216 (22.48)
190 (19.59)
49 (18.92)
55 (25) 124 (56.36)
216 (22.48) 529 (55.05)
218 (22.47) 562 (57.94)
62 (23.94) 148 (57.14)
0.9 126 (52.72) 113 (47.28) 61 (7.7)
62 (11)
<0.001 0.002
<0.001 0.5 0.003
0.001
0.08
<0.001
0.6
Note: Boldface indicates statistical significance (p<0.05). HS/GED, high school/General Educational Development test.
Treatment uptake and quit attempts are reported in Appendix A and Appendix Tables 1 4. Briefly, the interactions between education or income X treatment arm with respect to all secondary outcomes related to treatment uptake and quit attempts were not statistically significant. Smokers at each level
October 2018
of education and income who were randomized to proactive care were more likely to report using telephone counseling or combined therapy than those randomized to usual care. VA smoking cessation medication use showed a weaker response to the proactive care intervention and quit attempts did
510
Table 2. Sample Characteristics by Income Level Characteristics
Total, n (%) or median (SD)
$10,000 $20,000, n (%) or median (SD)
$20,001 $40,000, n (%) or median (SD)
$40,001 $60,000, n (%) or median (SD)
>$60,000, n (%) or median (SD)
429
769
734
305
193
223 (51.98) 206 (48.02)
396 (51.5) 373 (48.5)
385 (52.45) 349 (47.55)
159 (52.13) 146 (47.87)
105 (54.4) 88 (45.6)
57 (9.1)
61 (9.3)
60 (10.2)
184 (42.89) 169 (39.39) 34 (7.93) 42 (9.79) 398 (92.77) 105 (24.48)
450 (58.52) 227 (29.52) 42 (5.46) 50 (6.5) 745 (96.88) 287 (37.32)
489 (66.62) 167 (22.75) 44 (5.99) 34 (4.63) 687 (93.6) 420 (57.22)
208 (68.2) 60 (19.67) 21 (6.89) 16 (5.25) 284 (93.11) 208 (68.2)
139 (72.02) 32 (16.58) 15 (7.77) 7 (3.63) 186 (96.37) 166 (86.01)
163 (38.81) 178 (42.38) 79 (18.81)
274 (36.15) 313 (41.29) 171 (22.56)
262 (36.34) 296 (41.05) 163 (22.61)
95 (31.46) 142 (47.02) 65 (21.52)
72 (37.5) 81 (42.19) 39 (20.31)
60 (10)
59 (10.9)
p-value — 0.97
<0.001 <0.001
0.005 <0.001 0.5
<0.001 104 (24.36) 210 (49.18) 113 (26.46)
198 (26.12) 391 (51.58) 169 (22.3)
209 (28.67) 376 (51.58) 144 (19.75)
90 (29.61) 169 (55.59) 45 (14.8)
73 (38.02) 98 (51.04) 21 (10.94)
124 (30.69)
247 (34.26)
278 (39.71)
141 (48.12)
114 (60.96)
89 (22.03)
145 (20.11)
160 (22.86)
63 (21.5)
38 (20.32)
191 (47.28)
329 (45.63)
262 (37.43)
89 (30.38)
35 (18.72)
<0.001
(continued on next page)
Danan et al / Am J Prev Med 2018;55(4):506 516
All participants 2,430 Treatment group Usual care 1,268 (52.18) Proactive care 1,162 (47.82) Demographic characteristics Age, years 60 (9.8) Race White 1,470 (60.49) Black 655 (26.95) Hispanic 156 (6.42) Other 149 (6.13) Male gender 2,300 (94.65) Married 1,186 (48.81) Smoking behaviors Cigarettes per day 10 866 (36.19) 11 20 1,010 (42.21) 21 517 (21.6) Time to first cigarette, minutes 31 674 (27.97) 6 30 1,244 (51.62) <5 492 (20.41) Social and environmental pressures Home smoking rules Not allowed 904 (39.22) anywhere Allowed some 495 (21.48) places/times Allowed anywhere 906 (39.31)
<$10,000, n (%) or median (SD)
www.ajpmonline.org
October 2018
0.3
<0.001
511
not differ significantly by treatment arm in most SES levels.
111 (60.33) 178 (61.81) 390 (54.47) 218 (53.69)
399 (57.74) Note: Boldface indicates statistical significance (p<0.05).
42 (22.83) 65 (22.57) 171 (23.88) 98 (24.14)
150 (21.71)
31 (16.85) 45 (15.63) 155 (21.65) 90 (22.17)
142 (20.55)
80 (11.13) 218 (30.32) 150 (20.86) 151 (21) 120 (16.69) 58 (14.39) 88 (21.84) 83 (20.6) 88 (21.84) 86 (21.34)
131 (18.79) 196 (28.12) 160 (22.96) 138 (19.8) 72 (10.33)
54 (18.82) 111 (38.68) 62 (21.6) 42 (14.63) 18 (6.27)
50 (27.03) 68 (36.76) 32 (17.3) 28 (15.14) 7 (3.78)
DISCUSSION
Friends who smoke None 373 (16.28) Less than half 681 (29.73) About half 487 (21.26) More than half 447 (19.51) All 303 (13.23) People important to me want me to quit smoking Strongly 463 (20.26) disagree to neutral Somewhat 526 (23.02) agree Strongly agree 1,296 (56.72)
<$10,000, n (%) or median (SD) Total, n (%) or median (SD) Characteristics
Table 2. Sample Characteristics by Income Level (continued)
$10,000 $20,000, n (%) or median (SD)
$20,001 $40,000, n (%) or median (SD)
$40,001 $60,000, n (%) or median (SD)
>$60,000, n (%) or median (SD)
p-value
Danan et al / Am J Prev Med 2018;55(4):506 516
In a large, multicenter smoking-cessation trial, a proactive outreach intervention was associated with improved rates of prolonged abstinence among smokers at all SES levels. The largest effect sizes for the intervention were among smokers at the lowest SES levels, who are historically the least likely to quit successfully. A proactive approach to smoking cessation may both lower the overall population-level smoking rate, as shown in the primary outcomes paper,32 and reduce SES-related smoking disparities, as shown here. These findings address a gap in evidence related to pro-equity smoking cessation support interventions. Cessation support interventions (including counseling, medications, or combined therapy) can have a net-negative impact on equity because they are less effective among low-SES smokers.18,39 A reactive treatment approach and the financial cost of treatment, often associated with lower use of cessation support services40 and lower quit rates in low-SES populations, were overcome in this study by (1) using proactive outreach and telephone-based treatment and (2) facilitating access to free or low-cost cessation medications. Traditional smoking-cessation interventions rely on motivated smokers to seek out services. Smokers who independently call quitlines for support41 or volunteer for intervention programs42 are more likely to come from a higher educational background than the general smoking population. By proactively calling smokers, the proactive outreach intervention addressed the fear of being judged, fear of failure, and lack of knowledge43 that may prevent low-SES smokers from seeking assistance. In addition, this telephone-based intervention did not rely on smokers to interface with the healthcare system directly, potentially increasing utilization by lowSES smokers.44 The cost of cessation medications is also a barrier to seeking and using care,45 and most lowincome U.S. smokers are unaware of insurance-subsidized treatment benefits.46 Integrating the offer of free or low-cost cessation medications with proactive outreach increased awareness of their availability, which is associated with increased utilization.46 Several studies have used proactive outreach to recruit smokers to quit,47 54 though none have evaluated the differential effect of proactive outreach interventions on smokers at different SES levels. Tzelepis et al.55 evaluated the overall relationship between SES and prolonged abstinence among actively recruited smokers. They found only shortterm cessation success for the primary intervention,54 and
512
Danan et al / Am J Prev Med 2018;55(4):506 516
Figure 1. Model-based proportion of smokers who reported 6-month prolonged abstinence at 1-year follow-up, by (A) education or (B) income level and treatment group. GED, General Educational Development test.
reported that among all smokers, being employed predicted one abstinence outcome at one timepoint, but other SES measures (education and private insurance) were not associated with abstinence.55 They did not assess the interaction between treatment group and SES. A pooled analysis of the intervention groups from five proactively recruited smoking-cessation studies (combined n=2,972) found a small but significant positive effect of education on point-prevalent abstinence at 12 and 24 months’ follow-up.56 Qualitatively, a similar pattern was observed among smokers of different education levels within the proactive care arm of this trial, for all except the lowest education level. The current study is enriched by the comparison with abstinence rates across education levels in the usual-care group arm. Proactive outreach has recently been shown to decrease smoking in low-SES populations in studies that
did not include higher-SES smokers.57,58 A comparable effect size at all SES levels was expected in the present study because knowledge, access, and utilization of the intervention were not contingent on individual or community resources. The nonsignificant interaction term supported this hypothesis, but it is surprising to find that the largest effect sizes for the intervention were in the lowest SES groups. Relatively high observed prolonged abstinence rates of 17% (11th grade education or less) or nearly 19% (less than $10,000 annual income) were unanticipated given that, consistent with historical patterns,6,7 low-SES smokers were heavier smokers with lower social and environmental pressures to quit. Variation in the utilization of cessation treatments or number of quit attempts (the secondary outcomes analyzed) did not explain this result. www.ajpmonline.org
513
Danan et al / Am J Prev Med 2018;55(4):506 516
Table 3. Six-Month Prolonged Abstinence (n, Model-Based %) by Treatment Arm and SES Level, With Model-Based OR Participants, n (%) SES level Education n 11th grade, n=239 HS/GED, n=1,031 Some college, n=1,016 College, n=279 Income, USD n <$10,000, n=429 $10,000 $20,000, n=769 $20,001 $40,000, n=734 $40,001 $60,000, n=305 >$60,000, n=193
Usual care
Proactive care
Complete casea model, OR (95% CI)
NMARb model, OR (95% CI)
1,342 6 (5.7) 47 (10.8) 36 (8.0) 12 (9.9)
1,223 16 (17.3) 46 (12.0) 60 (14.6) 23 (18.6)
2,565 3.46 (1.39, 8.59) 1.13 (0.76, 1.68) 1.96 (1.30, 2.96) 2.09 (1.03, 4.27)
5,123 2.05 (1.30, 3.24) 1.12 (0.91, 1.38) 1.42 (1.11, 1.82) 1.57 (0.87, 2.82)
1,268 18 (9.4) 29 (8.7) 31 (9.8) 9 (7.5) 12 (13.5)
1,162 32 (18.7) 36 (11.7) 45 (15.2) 11 (11.9) 12 (15.9)
2,430 2.23 (1.24, 3.98) 1.40 (0.86, 2.28) 1.65 (1.04, 2.62) 1.65 (0.74, 2.69) 1.21 (0.53, 2.78)
5,123 1.61 (1.09, 2.39) 1.17 (0.89, 1.54) 1.30 (1.00, 1.69) 1.54 (0.92, 2.56) 1.25 (0.70, 2.22)
Note: Boldface indicates statistical significance (p<0.05). a Complete case model: Usual care is reference group. Model controls for site, treatment arm, interaction between SES and treatment arm, gender, and age. Complete case includes follow-up survey responders for whom quit status is known. b NMAR model: Usual care is reference group. Likelihood-based not missing at random (NMAR) model accounts for non-response. HS/GED, high school/General Educational Development test.
Other possible explanations merit exploration. First, there may be pent-up demand for cessation assistance among low-SES smokers that leads to a preferential response to proactive outreach. Low-SES smokers making a first or second supported quit attempt may be more successful than high-SES smokers who have failed many previous quit attempts. One quitline evaluation found a stronger relationship between nicotine replacement therapy use and abstinence among low-SES smokers compared with high-SES smokers.43 On the other hand, higher-SES individuals who continue to smoke despite lower levels of nicotine dependence, high social and environmental pressures to quit, and many available quit resources may be resistant to additional outreach. Additionally, recent work by Hammett and colleagues59 confirmed that low-SES smokers often experience lower levels of perceived smoking-related stigma, or the social unacceptability of smoking. Low-SES smokers with a lower level of perceived smoking-related stigma at baseline reported higher abstinence rates at 1-year in response to a proactive outreach intervention than those with higher baseline levels of smoking-related stigma. It was hypothesized that proactive outreach contributed to denormalization of smoking among low-SES smokers whose social environments were conducive to tobacco use. In this trial, capturing smoking status in the electronic medical record made it possible to proactively reach out to all smokers, regardless of intent to quit, to facilitate cessation assistance. Although the VA healthcare system October 2018
was one of the first to adopt this practice, it is a recommended component of system-level change for smoking cessation that is becoming widespread.60 This strategy helped connect vulnerable subpopulations of smokers to evidence-based cessation treatments. Future trials of outreach to smokers, which are increasingly relying on mobile health technology,61,62 interactive voice response,57 and text messaging,61,63 65 should evaluate how those technologies address the needs of smokers at all SES levels.59
Limitations Prolonged abstinence is a self-reported measure that may be subject to social desirability bias. However, multiple studies have validated smoking self-report,66,67 including among VA smokers,37 and most research has failed to identify systematic differences in the risk of misclassification of smokers by SES level.68,69 Interaction tests may be underpowered, and the original trial was not designed to detect subgroup effects by SES level. The population of mostly male U.S. Veterans limits generalizability to female Veterans and non-Veterans. Finally, SES is a complex construct that is only partially captured by measured variables, such as education and income levels. Race, class, occupation, and subjective social standing all contribute to SES, but are outside the scope of this paper. For an analysis of the effect of proactive outreach on prolonged abstinence among African American and white smokers in the Victory trial, please see Burgess et al.70
514
Danan et al / Am J Prev Med 2018;55(4):506 516
CONCLUSIONS Future tobacco control policies and interventions must promote equity. Proactive outreach interventions have the potential to decrease population-level smoking prevalence while reducing smoking-related disparities.
ACKNOWLEDGMENTS The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the U.S. government. This study was funded by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, and Health Services Research and Development (IAB-05-303). ED collaborated on the conceptualization of this study, guided the analytic approach, interpreted the results, and wrote most sections of the draft manuscript. She oversaw all aspects of this study and takes full and final responsibility for the paper. SF secured funding for and designed the original Victory trial and collaborated on the conceptualization of the current analysis. SN and BC designed and carried out the statistical analysis and assisted with interpretation and wrote sections of the methods and results. RW and PH collaborated on the conceptualization of this study, critically revised the manuscript, and read and approved the final submitted version. DB is the senior author on the paper; she led the conceptualization and helped design the analytic approach, interpret the study findings, and revise the manuscript. All authors critically read, provided feedback, and edited the final paper. Registered in clinicaltrials.gov (NCT00608426). Earlier versions of this study were presented in a poster presentation at the Society for Research on Nicotine and Tobacco (Florence, Italy, March 2017); the Society for General Internal Medicine (Washington, DC, April 2017); and the Veterans Affairs Health Services Research and Development National Conference (Washington, DC, July 2017). No financial disclosures were reported by the authors of this paper.
SUPPLEMENTAL MATERIAL Supplemental materials associated with this article can be found in the online version at doi:10.1016/j. amepre.2018.05.023.
REFERENCES 1. Lopez AD, Collishaw NE, Piha T. A descriptive model of the cigarette epidemic in developed countries. Tob Control. 1994;3(3):242–247. https://doi.org/10.1136/tc.3.3.242. 2. Thun M, Peto R, Boreham J, Lopez AD. Stages of the cigarette epidemic on entering its second century. Tob Control. 2012;21(2):96–101. https://doi.org/10.1136/tobaccocontrol-2011-050294. 3. HHS. The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta, GA: HHS, CDC, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Published 2014.
4. Whyte B, Gordon D, Haw S, Fischbacher C, Scott N, Harrison R. An Atlas of Tobacco Smoking in Scotland: A Report Presenting Estimated Smoking Prevalence and Smoking Attributable Deaths Within Scotland. NHS Health Scotland. Published 2007. 5. Kunst A, Giskes K, Mackenbach J. Socio-economic Inequalities in Smoking in the European Union: Applying an Equity Lens to Tobacco Control Policies: for the EU Network on Interventions to Reduce Socioeconomic Inequalities in Health. Erasmus Medical Center Rotterdam, 2004. 6. Siahpush M, McNeill A, Borland R, Fong GT. Socioeconomic variations in nicotine dependence, self-efficacy, and intention to quit across four countries: findings from the International Tobacco Control (ITC) Four Country Survey. Tob Control. 2006;15(suppl 3):iii71–iii75. https://doi.org/10.1136/tc.2004.008763. 7. Kotz D, West R. Explaining the social gradient in smoking cessation: it’s not in the trying, but in the succeeding. Tob Control. 2009;18 (1):43–46. https://doi.org/10.1136/tc.2008.025981. 8. Mackenbach JP, Huisman M, Andersen O, et al. Inequalities in lung cancer mortality by the educational level in 10 European populations. Eur J Cancer. 2004;40(1):126–135. https://doi.org/10.1016/j. ejca.2003.10.018. 9. Jha P, Peto R, Zatonski W, Boreham J, Jarvis MJ, Lopez AD. Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet. 2006;368(9533):367–370. https:// doi.org/10.1016/S0140-6736(06)68975-7. 10. Gregoraci G, van Lenthe FJ, Artnik B, et al. Contribution of smoking to socioeconomic inequalities in mortality: a study of 14 European countries, 1990 2004. Tob Control. 2017;26(3):260–268. https://doi. org/10.1136/tobaccocontrol-2015-052766. 11. Link BG, Phelan J. Social conditions as fundamental causes of disease. J Health Soc Behav. 1995:80 94. https://doi.org/10.2307/2626958. 12. Phelan JC, Link BG. Controlling disease and creating disparities: a fundamental cause perspective. J Gerontol B Psychol Sci Soc Sci. 2005;60(Special Issue 2):S27–S33. 13. Bauld L, Judge K, Platt S. Assessing the impact of smoking cessation services on reducing health inequalities in England: observational study. Tob Control. 2007;16(6):400–404. https://doi.org/10.1136/ tc.2007.021626. 14. Browning KK, Ferketich AK, Salsberry PJ, Wewers ME. Socioeconomic disparity in provider-delivered assistance to quit smoking. Nicotine Tob Res. 2008;10(1):55–61. https://doi.org/10.1080/14622200701704905. 15. Niederdeppe J, Fiore MC, Baker TB, Smith SS. Smoking-cessation media campaigns and their effectiveness among socioeconomically advantaged and disadvantaged populations. Am J Public Health. 2008;98(5):916–924. https://doi.org/10.2105/AJPH.2007.117499. 16. Jarvis MJ, Wardle J. Social Patterning of Individual Health Behaviours: The Case of Cigarette Smoking. Oxford, UK: Oxford University Press; 2005. https://doi.org/10.1093/acprof:oso/9780198565895.003.11. 17. Thomas S, Fayter D, Misso K, et al. Population tobacco control interventions and their effects on social inequalities in smoking: systematic review. Tob Control. 2008;17(4):230–237. https://doi.org/10.1136/ tc.2007.023911. 18. Hill S, Amos A, Clifford D, Platt S. Impact of tobacco control interventions on socioeconomic inequalities in smoking: review of the evidence. Tob Control. 2014;23(e2):e89–e97. https://doi.org/10.1136/ tobaccocontrol-2013-051110. 19. Remler DK. Poor smokers, poor quitters, and cigarette tax regressivity. Am J Public Health. 2004;94(2):225–229. https://doi.org/10.2105/ AJPH.94.2.225. 20. West R, McEwen A, Bolling K, Owen L. Smoking cessation and smoking patterns in the general population: a 1-year follow-up. Addiction. 2001;96 (6):891–902. https://doi.org/10.1046/j.1360-0443.2001.96689110.x. 21. Vangeli E, Stapleton J, Smit ES, Borland R, West R. Predictors of attempts to stop smoking and their success in adult general
www.ajpmonline.org
Danan et al / Am J Prev Med 2018;55(4):506 516
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
population samples: a systematic review. Addiction. 2011;106 (12):2110–2121. https://doi.org/10.1111/j.1360-0443.2011.03565.x. Chandola T, Head J, Bartley M. Socio-demographic predictors of quitting smoking: how important are household factors. Addiction. 2004;99 (6):770–777. https://doi.org/10.1111/j.1360-0443.2004.00756.x. Giskes K, van Lenthe FJ, Turrell G, Brug J, Mackenbach JP. Smokers living in deprived areas are less likely to quit: a longitudinal follow-up. Tob Control. 2006;15(6):485–488. https://doi.org/10.1136/tc.2006.015750. Lacey LP, Manfredi C, Balch G, Warnecke RB, Allen K, Edwards C. Social support in smoking cessation among black women in Chicago public housing. Public Health Rep. 1993;108(3):387–394. Mermelstein R, Cohen S, Lichtenstein E, Baer JS, Kamarck T. Social support and smoking cessation and maintenance. J Consult Clin Psychol. 1986;54(4):447–453. https://doi.org/10.1037/0022006X.54.4.447. Businelle MS, Kendzor DE, Reitzel LR, et al. Mechanisms linking socioeconomic status to smoking cessation: a structural equation modeling approach. Health Psychol. 2010;29(3):262–273. https://doi. org/10.1037/a0019285. Lam T-H, Abdullah ASM, Chan SS, Hedley AJ. Adherence to nicotine replacement therapy versus quitting smoking among Chinese smokers: a preliminary investigation. Psychopharmacology. 2005;177 (4):400–408. https://doi.org/10.1007/s00213-004-1971-y. Nevid JS, Javier RA, Moulton III JL. Factors predicting participant attrition in a community-based, culturally specific smokingcessation program for Hispanic smokers. Health Psychol. 1996;15 (3):226–229. https://doi.org/10.1037/0278-6133.15.3.226. Hiscock R, Judge K, Bauld L. Social inequalities in quitting smoking: what factors mediate the relationship between socioeconomic position and smoking cessation?. J Public Health (Oxf). 2011;33(1):39–47. https://doi.org/10.1093/pubmed/fdq097. Phelan JC, Link BG, Tehranifar P. Social conditions as fundamental causes of health inequalities: theory, evidence, and policy implications. J Health Soc Behav. 2010;51(1):S28–S40. (suppl) https://doi.org/ 10.1177/0022146510383498. Fu SS, van Ryn M, Sherman SE, et al. Population-based tobacco treatment: study design of a randomized controlled trial. BMC Public Health. 2012;12:159 https://doi.org/10.1186/1471-2458-12-159. Fu SS, van Ryn M, Sherman SE, et al. Proactive tobacco treatment and population-level cessation: a pragmatic randomized clinical trial. JAMA Intern Med. 2014;174(5):671–677. https://doi.org/10.1001/ jamainternmed.2014.177. Tzelepis F, Paul CL, Walsh RA, McElduff P, Knight J. Proactive telephone counseling for smoking cessation: meta-analyses by recruitment channel and methodological quality. J Natl Cancer Inst. 2011;103(12):922–941. https://doi.org/10.1093/jnci/djr169. Zwarenstein M, Treweek S, Gagnier JJ, et al. Improving the reporting of pragmatic trials: an extension of the CONSORT statement. BMJ. 2008;337:a2390 https://doi.org/10.1136/bmj.a2390. Hughes JR, Keely JP, Niaura RS, Ossip-Klein DJ, Richmond RL, Swan GE. Measures of abstinence in clinical trials: issues and recommendations. Nicotine Tob Res. 2003;5(1):13–25. https://doi.org/10.1080/ 1462220031000070552. Benowitz NL, Jacob III P, Ahijevych K, et al. Biochemical verification of tobacco use and cessation. Nicotine Tob Res. 2002;4(2):149–159. https://doi.org/10.1080/14622200210123581. Noonan D, Jiang Y, Duffy SA. Utility of biochemical verification of tobacco cessation in the Department of Veterans Affairs. Addict Behav. 2013;38(3):1792–1795. https://doi.org/10.1016/j.addbeh.2012.11.006. Ibrahim JG, Chen MH, Lipsitz SR. Missing responses in generalised linear mixed models when the missing data mechanism is nonignorable. Biometrika. 2001;88(2):551–564. https://doi.org/10.1093/biomet/88.2.551. Main C, Thomas S, Ogilvie D, et al. Population tobacco control interventions and their effects on social inequalities in smoking: placing an
October 2018
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
515
equity lens on existing systematic reviews. BMC Public Health. 2008;8:178. https://doi.org/10.1186/1471-2458-8-178. Roddy E, Antoniak M, Britton J, Molyneux A, Lewis S. Barriers and motivators to gaining access to smoking cessation services amongst deprived smokers—a qualitative study. BMC Health Serv Res. 2006;6:147 https://doi.org/10.1186/1472-6963-6-147. Tzelepis F, Paul CL, Walsh RA, et al. Telephone recruitment into a randomized controlled trial of quitline support. Am J Prev Med. 2009;37(4):324–329. https://doi.org/10.1016/j.amepre.2009.05.022. Monso E, Campbell J, Tønnesen P, Gustavsson G, Morera J. Sociodemographic predictors of success in smoking intervention. Tob Control. 2001;10(2):165–169. https://doi.org/10.1136/tc.10.2.165. Murray RL, Bauld L, Hackshaw LE, McNeill A. Improving access to smoking cessation services for disadvantaged groups: a systematic review. J Public Health. 2009;31(2):258–277. https://doi.org/10.1093/ pubmed/fdp008. Bryant J, Bonevski B, Paul C, McElduff P, Attia J. A systematic review and meta-analysis of the effectiveness of behavioural smoking cessation interventions in selected disadvantaged groups. Addiction. 2011;106 (9):1568–1585. https://doi.org/10.1111/j.1360-0443.2011.03467.x. Beauchamp A, Peeters A, Tonkin A, Turrell G. Best practice for prevention and treatment of cardiovascular disease through an equity lens: a review. Eur J Cardiovasc Prev Rehabil. 2010;17(5):599–606. https://doi.org/10.1097/HJR.0b013e328339cc99. McMenamin SB, Halpin HA, Bellows NM. Knowledge of Medicaid coverage and effectiveness of smoking treatments. Am J Prev Med. 2006;31(5):369–374. https://doi.org/10.1016/j.amepre.2006.07.015. Abdullah ASM, Mak YW, Loke AY, Lam TH. Smoking cessation intervention in parents of young children: a randomised controlled trial. Addiction. 2005;100(11):1731–1740. https://doi.org/10.1111/ j.1360-0443.2005.01231.x. Aveyard P, Griffin C, Lawrence T, Cheng K. A controlled trial of an expert system and self-help manual intervention based on the stages of change versus standard self-help materials in smoking cessation. Addiction. 2003;98(3):345 https://doi.org/10.1046/j.1360-0443.2003.00302.x. McBride CM, Scholes D, Grothaus LC, Curry SJ, Ludman E, Albright J. Evaluation of a minimal self-help smoking cessation intervention following cervical cancer screening. Prev Med. 1999;29(2):133–138. https://doi.org/10.1006/pmed.1999.0514. Curry SJ, McBride C, Grothaus LC, Louie D, Wagner EH. A randomized trial of self-help materials, personalized feedback, and telephone counseling with nonvolunteer smokers. J Consult Clin Psychol. 1995;63(6):1005 https://doi.org/10.1037/0022-006X.63.6.1005. Lichtenstein E, Boles SM, Lee ME, Hampson SE, Glasgow RE, Fellows J. Using radon risk to motivate smoking reduction II: randomized evaluation of brief telephone counseling and a targeted video. Health Educ Res. 2007;23(2):191–201. https://doi.org/10.1093/her/cym016. Lichtenstein E, Andrews JA, Lee ME, Glasgow RE, Hampson SE. Using radon risk to motivate smoking reduction: evaluation of written materials and brief telephone counselling. Tob Control. 2000;9(3):320– 326. https://doi.org/10.1136/tc.9.3.320. Prochaska JO, Velicer WF, Fava JL, et al. Counselor and stimulus control enhancements of a stage-matched expert system intervention for smokers in a managed care setting. Prev Med. 2001;32(1):23–32. https://doi.org/10.1006/pmed.2000.0767. Tzelepis F, Paul CL, Wiggers J, et al. A randomised controlled trial of proactive telephone counselling on cold-called smokers’ cessation rates. Tob Control. 2011;20(1):40–46. https://doi.org/10.1136/tc.2010.035956. Tzelepis F, Paul CL, Walsh RA, Wiggers J, Duncan SL, Knight J. Predictors of abstinence among smokers recruited actively to quitline support. Addiction. 2013;108(1):181–185. https://doi.org/10.1111/j.13600443.2012.03998.x. Velicer WF, Redding CA, Sun X, Prochaska JO. Demographic variables, smoking variables, and outcome across five studies. Health Psychol. 2007;26(3):278 https://doi.org/10.1037/0278-6133.26.3.278.
516
Danan et al / Am J Prev Med 2018;55(4):506 516
57. Haas JS, Linder JA, Park ER, et al. Proactive tobacco cessation outreach to smokers of low socioeconomic status: a randomized clinical trial. JAMA Intern Med. 2015;175(2):218–226. https://doi.org/10.1001/ jamainternmed.2014.6674. 58. Fu SS, van Ryn M, Nelson D, et al. Proactive tobacco treatment offering free nicotine replacement therapy and telephone counselling for socioeconomically disadvantaged smokers: a randomised clinical trial. Thorax. 2016;71(5):446–453. https://doi.org/10.1136/thoraxjnl-2015-207904. 59. Hammett P, Fu SS, Nelson D, et al. A proactive smoking cessation intervention for socioeconomically disadvantaged smokers: the role of smoking-related stigma. Nicotine Tob Res. 2018;20(3):286–294. https://doi.org/10.1093/ntr/ntx085. 60. Thomas D, Abramson MJ, Bonevski B, George J. System change interventions for smoking cessation. Cochrane Database Syst Rev. 2017;2: CD010742 https://doi.org/10.1002/14651858.CD010742.pub2. 61. Free C, Phillips G, Galli L, et al. The effectiveness of mobile-health technology-based health behaviour change or disease management interventions for health care consumers: a systematic review. PLoS Med. 2013;10 (1):e1001362 https://doi.org/10.1371/journal.pmed.1001362. 62. Heffner JL, Mull KE. Smartphone ownership among U.S. adult cigarette smokers: 2014 Health Information National Trends Survey (HINTS) data. J Med Internet Res. 2017;19(8):e305 https://doi.org/10.2196/jmir.7953. 63. Whittaker R, Borland R, Bullen C, Lin RB, McRobbie H, Rodgers A. Mobile phone-based interventions for smoking cessation. Cochrane Database Syst Rev. 2009;4:CD006611 https://doi.org/10.1002/14651858.CD006611.pub2.
64. Armanasco AA, Miller YD, Fjeldsoe BS, Marshall AL. Preventive health behavior change text message interventions: a meta-analysis. Am J Prev Med. 2017;52(3):391–402. https://doi.org/10.1016/j. amepre.2016.10.042. 65. Abroms LC, Johnson PR, Leavitt LE, et al. A randomized trial of text messaging for smoking cessation in pregnant women. Am J Prev Med. 2017;53(6):781–790. https://doi.org/10.1016/j.amepre.2017.08.002. 66. Yeager DS, Krosnick JA. The validity of self-reported nicotine product use in the 2001 2008 National Health and Nutrition Examination Survey. Med Care. 2010;48(12):1128–1132. https://doi.org/10.1097/ MLR.0b013e3181ef9948. 67. Patrick DL, Cheadle A, Thompson DC, Diehr P, Koepsell T, Kinne S. The validity of self-reported smoking: a review and meta-analysis. Am J Public Health. 1994;84(7):1086–1093. https://doi.org/10.2105/AJPH.84.7.1086. 68. Vartiainen E, Sepp€al€a T, Lillsunde P, Puska P. Validation of self reported smoking by serum cotinine measurement in a communitybased study. J Epidemiol Community Health. 2002;56(3):167–170. https://doi.org/10.1136/jech.56.3.167. 69. Graham H, Owen L. Are there socioeconomic differentials in underreporting of smoking in pregnancy. Tob Control. 2003;12(4):434 https://doi.org/10.1136/tc.12.4.434. 70. Burgess DJ, Van Ryn M, Noorbaloochi S, et al. Smoking cessation among African American and white smokers in the Veterans Affairs health care system. Am J Public Health. 2014;104(suppl 4):S580–S587. https://doi.org/10.2105/AJPH.2014.302023.
www.ajpmonline.org