Nicotine & Tobacco Research, Volume 16, Number 6 (June 2014) 689–696
Original Investigation
The Potential and Peril of Health Insurance Tobacco Surcharge Programs: Evidence From Georgia’s State Employees’ Health Benefit Plan Alex C. Liber MSPH1,2, Jason M. Hockenberry PhD1,3, Laura M. Gaydos PhD1, Joseph Lipscomb PhD1,4 1Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, GA; 2Economic and Health Policy Research, American Cancer Society, Atlanta, GA; 3Center for Comprehensive Access Delivery Research and Evaluation, Iowa City VA Healthcare System, Iowa City, IA; 4Winship Cancer Institute, Emory University, Atlanta, GA
Corresponding Author: Alex C. Liber, MSPH, International Tobacco Control Research, American Cancer Society, 250 Williams Street NW, Suite 6000, Atlanta, GA 30303, USA. Fax: 404-327-6450; E-mail: [email protected] Received July 15, 2013; accepted November 26, 2013
Introduction: A rapidly growing number of U.S. employers are charging health insurance surcharges for tobacco use to their employees. Despite their potential to price-discriminate, little systematic empirical evidence of the impacts of these tobacco surcharges has been published. We attempted to assess the impact of a health insurance surcharge for tobacco use on cessation among enrollees in Georgia’s State Health Benefit Plan (GSHBP). Methods: We identified a group of enrollees in GSHBP who began paying the tobacco surcharge at the program’s inception in July 2005. We examined the proportion of these enrollees who certified themselves and their family members as tobacco-free and no longer paid the surcharge through April 2011, and we defined this as implied cessation. We compared this proportion to a national expected annual 2.6% cessation rate. We also compared our observation group to a comparison group to assess surcharge avoidance. Results: By April 2011, 45% of enrollees who paid a tobacco surcharge starting in July 2005 had certified themselves as tobacco-free. This proportion exceeded the expected cessation based on 3 times the national rate (p < .001). The length of enrollment was not statistically different between our observation and comparison groups (p = .427). Conclusions: The reported rates of tobacco cessation among GSHBP enrollees resulting from a tobacco surcharge substantially exceed national rates. These surcharges appear to be effective, but the value of these results, and the effectiveness of health insurance surcharges in changing behavior, are tempered by the important limitation that enrollees’ certification of quitting was self-reported and not subject to additional, clinical verification.
Introduction The decline in the prevalence of regular tobacco use, mainly in the form of cigarette smoking, among adults in the United States has leveled off in the last decade (CDC’s Office on Smoking and Health, 2011). Separately and simultaneously, average health insurance premiums in the United States have risen precipitously (KFF, 2007). More recently, an increasing number of employers are requiring employees who use tobacco to pay a surcharge on their health insurance premium (Abelson, 2011; Towers Watson, 2010). Assessing a surcharge on the health insurance premium of employees who use tobacco serves at least three purposes from the employer’s perspective: First, charging an additional health insurance premium allocates some of the additional expected healthcare costs of smoking (which will raise premiums) to workers who smoke rather than spreading the cost
across the entire workforce. Second, raising the premium of tobacco users, assuming tobacco use is verifiable and the surcharge enforceable, financially incents quitting, and inducing quitting can reduce health care costs within a short time frame (Hockenberry et al., 2012). Third, raising the premium tobacco users pay may discourage them from actively seeking employment with the firm, or cause those who are currently employed (and unable to quit) to seek health insurance coverage and/ or employment elsewhere. Despite the growing number of employers implementing surcharges, little empirical evidence of the impacts of these tobacco surcharges on behavior has been published (Penner, 1989; Volk & Corlette, 2012). As of 2013, 12 state governments assessed tobacco use surcharges for their employees (Cauchi & Marsh, 2013). In July 2005, the Georgia State Health Benefit Plan (GSHBP) instituted a $40 per month tobacco use surcharge on health insurance. The tobacco surcharge policy included the following elements:
doi:10.1093/ntr/ntt216 Advance Access publication December 27, 2013 © The Author 2013. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected].
689
Downloaded from http://ntr.oxfordjournals.org/ at University of Pittsburgh on March 30, 2015
Abstract
Potential and peril of health insurance tobacco surcharge programs
Methods Data The data for this study are monthly health insurance enrollment data from July 2005 to April 2011 for enrollees in the GSHBP for every enrollee who paid the tobacco surcharge. The GSHBP is managed by the Georgia Department of Community Health (DCH), which provided the original dataset (non-identifiable to specific individuals) to the authors. The DCH did not provide access to the data of those enrollees who never paid the surcharge. The primary measure used in this study is an implied tobacco cessation event derived from the termination of payment of a monthly tobacco surcharge by a GSHBP enrollee,
Original Sample Ever Paid the Tobacco Surcharge n=102,137
Included Excluded
Never Enrolled in Medicare or Retired During the Study Period n=89,419
Paid Surcharge at First Observation n=75,112
Observation Group Enrolled in GSHBP in July 2005 n=43,034
which is potentially a proxy for tobacco cessation. When the authors refer to individual enrollee characteristics, we are describing the state employees that are the primary enrollees on GSHBP insurance policies. For the purposes of our analysis, we assume that all primary enrollees, no matter if they are the heads of families with children, are married, or are single, carry equal statistical weight in our calculations of implied cessation rates and population demographics. The enrollment data from DCH indicated in which months an enrollee paid the tobacco surcharge during the study period. If an enrollee stopped paying the surcharge for some time but resumed paying later, we considered them to have continued using tobacco during that period. We do this to provide a conservative estimate of cessation behavior. Ceasing and resuming payment of the surcharge potentially indicates a cessation event followed by relapse, but it could also indicate a tobacco-using household member moving off of, and then back onto, the plan. Similarly, we carried forward the last observed tobacco surcharge payment status of enrollees who left the GSHBP during the study period to the end of the study period (April 2011) because disenrollment could represent surcharge avoidance. Our observation group is a cohort of persons who never retired or enrolled in Medicare during the study as well as were enrolled in the GSHBP and paid the tobacco surcharge in July 2005. The sample selection process is illustrated in Figure 1. DCH provided monthly observations for the 102,137 enrollees (about one-seventh of all GSHBP enrollees) who ever paid the tobacco surcharge from the policy’s introduction in July 2005 through the end of our study period in April 2011. From the original sample, we excluded 12,988 enrollees who either enrolled in Medicare or retired during the study period. From the remaining 89,419 enrollees, we set aside 27,025 enrollees who did not pay the tobacco surcharge at their first observation in our dataset as the pool from which to construct our comparison group to examine whether being a tobacco user likely caused individuals to leave the GSHBP. Among these there were 4,215 enrollees who were enrolled in the GSHBP at the start of the program in July 2005 that paid the surcharge
Not Enrolled in GSHBP in July 2005 n=19,090
Ever Enrolled in Medicare or Retired During the Study Period n=12,988
Did Not Pay Surcharge at First Observation n=27,025
Comparison Group Paid Surcharge 3 Months or Less and Enrolled in GSHBP in July 2005 n=4,215
Paid Surcharge More Than 3 Months or Entered GSHBP After July 2005 n=22,810
Figure 1. Sample selection process of Georgia State Health Benefit Plan enrollees who ever paid the tobacco surcharge.
690
Downloaded from http://ntr.oxfordjournals.org/ at University of Pittsburgh on March 30, 2015
(a) The surcharge would be imposed unless an enrollee affirmed during an online enrollment process that no enrolled family member (including the employee) had used tobacco products in the prior 12 months, (b) the primary enrollee (the person employed by the state government) was warned they would face the loss of health insurance coverage and/or a large monetary fine for falsifying tobacco use status, and (c) the primary enrollee was notified that the tobacco surcharge could be removed prospectively by completing a tobacco cessation program and signing an affidavit stating that no enrolled family member had used tobacco in the past 60 days (Badertscher & Salzer, 2005; Georgia Department of Community Health, 2007). This surcharge was raised to $60 per month in 2010, and $80 per month in 2011 (Georgia State Health Benefit Plan [GSHBP], 2009, 2011). GSHBP enrollees paid 25% of the total cost of their health insurance policies, which in July 2005 averaged about $150 per month before the surcharge (GSHBP, 2009, 2011). The authors, in a manner equivalent to the GSHBP’s method used to assess the tobacco surcharge, studied the GSHBP population at the family level. We used longitudinal data on GSHBP enrollees’ tobacco surcharge payment and insurance enrollment to examine self-reported tobacco use in response to this tobacco surcharge on health insurance.
Nicotine & Tobacco Research
Statistical Analysis To assess whether the tobacco surcharge led to greater than expected reported cessation rates among GSHBP enrollees we employed a variety of analytic approaches. First, we calculated tobacco use rates at the beginning of each plan year and the end of the study for our observation group. The tobacco use rate is the proportion of the observation group who had not yet had the tobacco surcharge removed at a given time during the study. The proportion of persons who had the tobacco surcharge removed at a given time was considered to have ceased tobacco use. We then compared this reported tobacco use rate, using comparison of proportions testing, to the expected tobacco use rate resulting from a national cessation rate of 2.6% per year, derived from analysis of Méndez, Warner, and Courant (1998) of the National Health Interview Surveys from 1965 to 1993 and extrapolated to reflect the age distribution of the country in 2005 (Méndez & Warner, 2008). Finally, we conducted sensitivity analyses of the tobacco use rate among the observation group compared to various cessation rates in the population at large, to examine the robustness of the effect of the tobacco surcharge on implied cessation. We report annual instead of monthly implied cessation because the majority of cessation events took place in the month of January at the beginning of each plan year, which also coincides with annual employerbased health insurance plan changes.
Analysis of the data was performed using STATA 12.0 (StataCorp, 2011). The study was approved by the Emory University IRB.
Results Descriptive characteristics of the observation and comparison groups are shown in Table 1. Significant differences, found using two-sided t-tests, included the fact that a smaller proportion of the observation group than the comparison group lived in the Atlanta metro area (41.6% vs. 48.6%) and a larger percentage were parents (64.4% vs. 48.7%). Based on failure to certify as tobacco-free in July 2005, the prevalence of tobacco use among enrollees in the GSHBP was between 8.6% and 14.4%. The uncertainty in the prevalence arises due to the potential of multiple smokers in the household. The low rate assumes that only one person in each plan used tobacco, while the high rate assumes all adults on a plan used tobacco. The tobacco use rate among college-educated Georgians was 10.7% for women and 11.7% for men, a group that overlaps greatly with the mostly female and better-educated state employee enrollees covered by the GSHBP (Bryan, Thompson, & Patel, 2006). The proportion of the observation group that continued to pay the tobacco use surcharge over the course of the study period (with CI bars) is shown in Figure 2A. We also plot the expected tobacco use rate based on a national cessation rate of 2.6% per year for visual comparison in Figure 2A. By January 2006, only 6 months after the tobacco surcharge was implemented, about one-fifth of the observation group was no longer paying the tobacco use surcharge and at the end of the study period in April 2011, that figure had increased to 45.0%. Alternatively, in Figure 2B, we plot the reported proportion of tobacco users among persons in the observation group who did not report quitting tobacco until after January 2006 against multiples of the national cessation rate. By the end of the study period, 34.8% of enrollees in this subsample of the observation group had reported quitting tobacco, at a rate between two and three times the national cessation rate (both, p
Original Investigation
The Potential and Peril of Health Insurance Tobacco Surcharge Programs: Evidence From Georgia’s State Employees’ Health Benefit Plan Alex C. Liber MSPH1,2, Jason M. Hockenberry PhD1,3, Laura M. Gaydos PhD1, Joseph Lipscomb PhD1,4 1Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, GA; 2Economic and Health Policy Research, American Cancer Society, Atlanta, GA; 3Center for Comprehensive Access Delivery Research and Evaluation, Iowa City VA Healthcare System, Iowa City, IA; 4Winship Cancer Institute, Emory University, Atlanta, GA
Corresponding Author: Alex C. Liber, MSPH, International Tobacco Control Research, American Cancer Society, 250 Williams Street NW, Suite 6000, Atlanta, GA 30303, USA. Fax: 404-327-6450; E-mail: [email protected] Received July 15, 2013; accepted November 26, 2013
Introduction: A rapidly growing number of U.S. employers are charging health insurance surcharges for tobacco use to their employees. Despite their potential to price-discriminate, little systematic empirical evidence of the impacts of these tobacco surcharges has been published. We attempted to assess the impact of a health insurance surcharge for tobacco use on cessation among enrollees in Georgia’s State Health Benefit Plan (GSHBP). Methods: We identified a group of enrollees in GSHBP who began paying the tobacco surcharge at the program’s inception in July 2005. We examined the proportion of these enrollees who certified themselves and their family members as tobacco-free and no longer paid the surcharge through April 2011, and we defined this as implied cessation. We compared this proportion to a national expected annual 2.6% cessation rate. We also compared our observation group to a comparison group to assess surcharge avoidance. Results: By April 2011, 45% of enrollees who paid a tobacco surcharge starting in July 2005 had certified themselves as tobacco-free. This proportion exceeded the expected cessation based on 3 times the national rate (p < .001). The length of enrollment was not statistically different between our observation and comparison groups (p = .427). Conclusions: The reported rates of tobacco cessation among GSHBP enrollees resulting from a tobacco surcharge substantially exceed national rates. These surcharges appear to be effective, but the value of these results, and the effectiveness of health insurance surcharges in changing behavior, are tempered by the important limitation that enrollees’ certification of quitting was self-reported and not subject to additional, clinical verification.
Introduction The decline in the prevalence of regular tobacco use, mainly in the form of cigarette smoking, among adults in the United States has leveled off in the last decade (CDC’s Office on Smoking and Health, 2011). Separately and simultaneously, average health insurance premiums in the United States have risen precipitously (KFF, 2007). More recently, an increasing number of employers are requiring employees who use tobacco to pay a surcharge on their health insurance premium (Abelson, 2011; Towers Watson, 2010). Assessing a surcharge on the health insurance premium of employees who use tobacco serves at least three purposes from the employer’s perspective: First, charging an additional health insurance premium allocates some of the additional expected healthcare costs of smoking (which will raise premiums) to workers who smoke rather than spreading the cost
across the entire workforce. Second, raising the premium of tobacco users, assuming tobacco use is verifiable and the surcharge enforceable, financially incents quitting, and inducing quitting can reduce health care costs within a short time frame (Hockenberry et al., 2012). Third, raising the premium tobacco users pay may discourage them from actively seeking employment with the firm, or cause those who are currently employed (and unable to quit) to seek health insurance coverage and/ or employment elsewhere. Despite the growing number of employers implementing surcharges, little empirical evidence of the impacts of these tobacco surcharges on behavior has been published (Penner, 1989; Volk & Corlette, 2012). As of 2013, 12 state governments assessed tobacco use surcharges for their employees (Cauchi & Marsh, 2013). In July 2005, the Georgia State Health Benefit Plan (GSHBP) instituted a $40 per month tobacco use surcharge on health insurance. The tobacco surcharge policy included the following elements:
doi:10.1093/ntr/ntt216 Advance Access publication December 27, 2013 © The Author 2013. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected].
689
Downloaded from http://ntr.oxfordjournals.org/ at University of Pittsburgh on March 30, 2015
Abstract
Potential and peril of health insurance tobacco surcharge programs
Methods Data The data for this study are monthly health insurance enrollment data from July 2005 to April 2011 for enrollees in the GSHBP for every enrollee who paid the tobacco surcharge. The GSHBP is managed by the Georgia Department of Community Health (DCH), which provided the original dataset (non-identifiable to specific individuals) to the authors. The DCH did not provide access to the data of those enrollees who never paid the surcharge. The primary measure used in this study is an implied tobacco cessation event derived from the termination of payment of a monthly tobacco surcharge by a GSHBP enrollee,
Original Sample Ever Paid the Tobacco Surcharge n=102,137
Included Excluded
Never Enrolled in Medicare or Retired During the Study Period n=89,419
Paid Surcharge at First Observation n=75,112
Observation Group Enrolled in GSHBP in July 2005 n=43,034
which is potentially a proxy for tobacco cessation. When the authors refer to individual enrollee characteristics, we are describing the state employees that are the primary enrollees on GSHBP insurance policies. For the purposes of our analysis, we assume that all primary enrollees, no matter if they are the heads of families with children, are married, or are single, carry equal statistical weight in our calculations of implied cessation rates and population demographics. The enrollment data from DCH indicated in which months an enrollee paid the tobacco surcharge during the study period. If an enrollee stopped paying the surcharge for some time but resumed paying later, we considered them to have continued using tobacco during that period. We do this to provide a conservative estimate of cessation behavior. Ceasing and resuming payment of the surcharge potentially indicates a cessation event followed by relapse, but it could also indicate a tobacco-using household member moving off of, and then back onto, the plan. Similarly, we carried forward the last observed tobacco surcharge payment status of enrollees who left the GSHBP during the study period to the end of the study period (April 2011) because disenrollment could represent surcharge avoidance. Our observation group is a cohort of persons who never retired or enrolled in Medicare during the study as well as were enrolled in the GSHBP and paid the tobacco surcharge in July 2005. The sample selection process is illustrated in Figure 1. DCH provided monthly observations for the 102,137 enrollees (about one-seventh of all GSHBP enrollees) who ever paid the tobacco surcharge from the policy’s introduction in July 2005 through the end of our study period in April 2011. From the original sample, we excluded 12,988 enrollees who either enrolled in Medicare or retired during the study period. From the remaining 89,419 enrollees, we set aside 27,025 enrollees who did not pay the tobacco surcharge at their first observation in our dataset as the pool from which to construct our comparison group to examine whether being a tobacco user likely caused individuals to leave the GSHBP. Among these there were 4,215 enrollees who were enrolled in the GSHBP at the start of the program in July 2005 that paid the surcharge
Not Enrolled in GSHBP in July 2005 n=19,090
Ever Enrolled in Medicare or Retired During the Study Period n=12,988
Did Not Pay Surcharge at First Observation n=27,025
Comparison Group Paid Surcharge 3 Months or Less and Enrolled in GSHBP in July 2005 n=4,215
Paid Surcharge More Than 3 Months or Entered GSHBP After July 2005 n=22,810
Figure 1. Sample selection process of Georgia State Health Benefit Plan enrollees who ever paid the tobacco surcharge.
690
Downloaded from http://ntr.oxfordjournals.org/ at University of Pittsburgh on March 30, 2015
(a) The surcharge would be imposed unless an enrollee affirmed during an online enrollment process that no enrolled family member (including the employee) had used tobacco products in the prior 12 months, (b) the primary enrollee (the person employed by the state government) was warned they would face the loss of health insurance coverage and/or a large monetary fine for falsifying tobacco use status, and (c) the primary enrollee was notified that the tobacco surcharge could be removed prospectively by completing a tobacco cessation program and signing an affidavit stating that no enrolled family member had used tobacco in the past 60 days (Badertscher & Salzer, 2005; Georgia Department of Community Health, 2007). This surcharge was raised to $60 per month in 2010, and $80 per month in 2011 (Georgia State Health Benefit Plan [GSHBP], 2009, 2011). GSHBP enrollees paid 25% of the total cost of their health insurance policies, which in July 2005 averaged about $150 per month before the surcharge (GSHBP, 2009, 2011). The authors, in a manner equivalent to the GSHBP’s method used to assess the tobacco surcharge, studied the GSHBP population at the family level. We used longitudinal data on GSHBP enrollees’ tobacco surcharge payment and insurance enrollment to examine self-reported tobacco use in response to this tobacco surcharge on health insurance.
Nicotine & Tobacco Research
Statistical Analysis To assess whether the tobacco surcharge led to greater than expected reported cessation rates among GSHBP enrollees we employed a variety of analytic approaches. First, we calculated tobacco use rates at the beginning of each plan year and the end of the study for our observation group. The tobacco use rate is the proportion of the observation group who had not yet had the tobacco surcharge removed at a given time during the study. The proportion of persons who had the tobacco surcharge removed at a given time was considered to have ceased tobacco use. We then compared this reported tobacco use rate, using comparison of proportions testing, to the expected tobacco use rate resulting from a national cessation rate of 2.6% per year, derived from analysis of Méndez, Warner, and Courant (1998) of the National Health Interview Surveys from 1965 to 1993 and extrapolated to reflect the age distribution of the country in 2005 (Méndez & Warner, 2008). Finally, we conducted sensitivity analyses of the tobacco use rate among the observation group compared to various cessation rates in the population at large, to examine the robustness of the effect of the tobacco surcharge on implied cessation. We report annual instead of monthly implied cessation because the majority of cessation events took place in the month of January at the beginning of each plan year, which also coincides with annual employerbased health insurance plan changes.
Analysis of the data was performed using STATA 12.0 (StataCorp, 2011). The study was approved by the Emory University IRB.
Results Descriptive characteristics of the observation and comparison groups are shown in Table 1. Significant differences, found using two-sided t-tests, included the fact that a smaller proportion of the observation group than the comparison group lived in the Atlanta metro area (41.6% vs. 48.6%) and a larger percentage were parents (64.4% vs. 48.7%). Based on failure to certify as tobacco-free in July 2005, the prevalence of tobacco use among enrollees in the GSHBP was between 8.6% and 14.4%. The uncertainty in the prevalence arises due to the potential of multiple smokers in the household. The low rate assumes that only one person in each plan used tobacco, while the high rate assumes all adults on a plan used tobacco. The tobacco use rate among college-educated Georgians was 10.7% for women and 11.7% for men, a group that overlaps greatly with the mostly female and better-educated state employee enrollees covered by the GSHBP (Bryan, Thompson, & Patel, 2006). The proportion of the observation group that continued to pay the tobacco use surcharge over the course of the study period (with CI bars) is shown in Figure 2A. We also plot the expected tobacco use rate based on a national cessation rate of 2.6% per year for visual comparison in Figure 2A. By January 2006, only 6 months after the tobacco surcharge was implemented, about one-fifth of the observation group was no longer paying the tobacco use surcharge and at the end of the study period in April 2011, that figure had increased to 45.0%. Alternatively, in Figure 2B, we plot the reported proportion of tobacco users among persons in the observation group who did not report quitting tobacco until after January 2006 against multiples of the national cessation rate. By the end of the study period, 34.8% of enrollees in this subsample of the observation group had reported quitting tobacco, at a rate between two and three times the national cessation rate (both, p
