Introduction

Traumatic spinal cord injury (SCI) leads to increased risk of secondary health conditions and early mortality.1, 2 For these individuals, there is an increased risk of respiratory complications, which are a leading cause of mortality,3 especially among those with high level SCI. However, even with increased respiratory issues, rates of cigarette smoking in this population are high.4, 5 Smoking can exacerbate the respiratory complications,6 prolong the course of pressure ulcers,7, 8 contribute to other negative health outcomes like edema4 and urinary tract infections,9 and increase risk of mortality.10 Estimates of current cigarette smoking among persons with SCI range from 24 to 48%, but studies are limited by small sample sizes or samples solely drawn from clinical samples.4, 5, 11, 12

Research in the general population has found a lower socioeconomic status (SES) to be associated with increased smoking rates13 and greater severity of chronic obstructive pulmonary disease.14 Persons with lower SES are less likely to receive smoking cessation counseling.15 Those in rural areas, where health-care accessibility is generally decreased, were found to have a higher likelihood of smoking.16 Recently, we11 investigated other risk factors in relation to cigarette smoking among persons with SCI identified through a rehabilitation setting and found persons with lower SES were more likely to be current smokers.

Smoking cessation is particularly important in this vulnerable population but has not been assessed. Persons with SCI are often limited with respect to access to care17 and have motor and/or sensory impairment that may impact the effectiveness of cessation interventions that have established efficacy in non-SCI smokers.18 Saunders et al.11 reported that among current smokers with SCI, 77.2% had ever tried to quit. Among persons who had attempted to quit, 50% made an attempt in the past year, and 29.9% had sought professional help. However, this study was focused in a clinical participant sample identified through a rehabilitation hospital and did not assess factors associated with smoking cessation.

Purpose

Limited research exists on cigarette smoking among those with SCI, and the existing studies have been conducted solely using participant samples identified through clinical settings, like inpatient rehabilitation, which can often exclude individuals with limited access to appropriate care. We herein sought to expand on the existing literature while at the same time minimizing sampling bias. Using a population-based cohort of persons with SCI, we aimed to (i) assess the prevalence of smoking, (ii) evaluate attempts to quit smoking and (iii) identify the relationship of access to care and SES with smoking status and attempts to quit smoking.

Methods

Population

We identified participants through the South Carolina SCI Surveillance System Registry (SCISSR), a population-based registry of SCI occurring in the state each year. All nonfederal hospitals in South Carolina are mandated to report discharge data on all hospitalizations involving SCI to the State Budget and Control Board through uniform billing discharge data (UB-04). The SCISSR does not include information from military or veteran hospitals and therefore represents the civilian population of South Carolina. Data in the SCISSR are validated through verification of randomly selected medical charts and have been shown to be 99% accurate and complete.19 International Classification of Diseases, 9th Revision, Clinical Modification codes of 806 [.0-.9] and 952 [.0-.9] were used to identify incident cases of SCI as defined by the Centers for Disease Control and Prevention for surveillance of SCI.20 Duplicate admissions were eliminated using personal identifiers. Nonresidents of South Carolina were excluded, as were nonincident cases. Participants were recruited based on hospital discharge from 1998 to 2012, and were eligible if at the time of the study they: (i) were ⩾18 years, (ii) were ⩾1 year post injury and (iii) had traumatic SCI with residual effects (that is, did not have a complete recovery).

Procedures

Data were collected through mail-in survey between 2010 and 2013. Prospective participants were sent an introductory letter describing the study and alerting them to materials that were sent 4–6 weeks later. Second mailings and phone calls were made to nonresponders. Participants returning surveys with large sections of missing data (usually by pages sticking together) were called as a follow-up to obtain missing information. Participants were offered $50 remuneration. All study procedures were approved by an institutional review board. Of 2063 participants meeting eligibility criteria, 833 responded to the postage paid mail-in survey (40.4%). Nonresponse was primarily due to inability to confirm current addresses, and a lack of phone numbers to contact potential participants. Respondents were more likely to be female (P=0.0475) and more likely to be white (P=0.0107). Nonrespondents were more likely to have a diagnosis of unspecified level of injury (P=0.0192) There were no differences in age at injury (P=0.7928).

Measures

Measurement of smoking and predictive factors were identical to those used with a clinical participant sample by Saunders et al.11 Briefly, smoking behavior was assessed using items from the Behavioral Risk Factor Surveillance System (BRFSS): (i) have you smoked at least 100 cigarettes in your entire life?, (ii) do you smoke right now, (iii) on average, about how many cigarettes a day do you now smoke and (iv) if you don’t currently smoke but smoked at one time, about how long has it been since you last smoked cigarettes regularly?21 Quitting history was assessed via two items developed for the survey: (i) if you are a current smoker, have you ever tried to quit smoking, (ii) if yes, did you seek professional help (such as counseling, a quit-smoking program, or receive prescription medication), and two items from the Smoking Information Questionnaire:22 (i) how long has it been since you tried to quit; and (ii) since you started smoking regularly, what is the longest time you have gone without smoking. Participants were defined as never smokers (had never smoked 100 cigarettes), former smokers (had smoked 100 cigarettes but did not currently smoke) and current smokers.

SES was measured by annual household income (<$15 000, $15 000–24 999, $25 000–74 999, $75 000+). Level of education was categorized as: less than a high school degree, high school or associates degree, bachelor’s degree or higher.

Demographic characteristics included age, age at injury, years post injury, gender and race. Injury severity was based on established categories of: (i) C1–C4, non-ambulatory, (ii) C5–C8, non-ambulatory, (iii) non-cervical non-ambulatory or (iv) ambulatory.23

Health-care access was measured using several indices. From the BRFSS, participants were asked ‘do you have one person you think of as your personal doctor or health care provider?’ and ‘if no, do you have more than one?’ With respect to check-ups, ‘about how long has it been since you last visited a doctor for a routine check-up?’ was categorized as in the past year, or more than a year ago/never. Insurance status was assessed through: ‘was there any time in the past year when you were completely without any health plan or health insurance coverage?’ Lastly, participants were asked ‘what kind of place do you go to most often for your health care?’ which was categorized as (i) clinic/health center/health maintenance organization, (ii) hospital emergency room and (iii) hospital outpatient department/other.

Analysis

All analyses were conducted using SASv9.4. Descriptive statistics were generated for smoking and smoking cessation variables. We then assessed the bivariate relationships of demographic, SES and health-care access variables with: (i) smoking status, (ii) attempting to quit in the past year and (iii) sought professional help to quit. The analyses of attempts to quit and seeking professional help focused only on current smokers. Chi-square tests were used for the categorical variables and Wilcoxon–Mann–Whitney tests for the continuous variables (Kruskal–Wallis test was used for smoking status as it is a 3-level variable). As our outcome was a 3-level variable, we used logistic regression to distinguish which groups differed for each level.

Statement of ethics

We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteer were followed during the course of this research.

Results

The majority of the participants were white, non-Hispanic (60.4%), male (71.9%) with a mean age of 49.4 (16.8) years. Sixty-five percent of the participants reported being able to walk at all, 6.2% had a C1–C4 non-ambulatory, 12.0% C5–C8 non-ambulatory, and 16.9% non-cervical non-ambulatory. Average age at injury was 43.6 (17.2) and average years post injury was 5.8 (3.8). Of the participants, 35.3% were current, 29.5% former and 35.2% never smokers (Table 1). Of the current smokers, 75.4% had ever tried to quit, and of these, only 27.9% of those who had ever attempted to quit sought professional help.

Table 1 Smoking characteristics

Injury severity was related to current smoking, with those with less severe injuries reporting higher rates of current smoking (Table 2). Demographic variables were also found to be associated with smoking, with males (36.6%) and races other than black or white (42.5%) reporting higher rates of current smoking. Additionally, current smokers were younger at survey and at injury. Those with lower SES were more likely to be current smokers, as were those with less access to health care (no regular doctor, no check-up in the past year, time in the past year with no insurance, use emergency room most often for care).

Table 2 Bivariate analyses between smoking status (current, former, never) and participant characteristics

Among current smokers, only younger age and lack of insurance coverage were associated with attempts (any vs none) to quit smoking in the past year (Table 3). Among quit attempters, use of cessation assistance was associated with (Table 4): female gender, white race, older age and having insurance. Clinically, individuals in need of physical assistance, who identified as having a personal doctor, and who received a check-up in the past year, were each more likely to use cessation support.

Table 3 Relationship of demographics, SES and health-care access to trying to quit smoking in the past yeara
Table 4 Relationship of demographics, SES and health-care access to having sought professional help to quit smokinga

Discussion

This study aimed to describe the prevalence of smoking and associated correlates of smoking behavior within an SCI population. Unlike prior studies examining smoking among SCI populations, this analysis is based on a population-based cohort, derived from a statewide registry of incident cases, thereby minimizing sampling bias that has been common in the literature to date. The rate of current smoking in this population-based cohort (35.3%) was higher than a recent report from a clinical cohort with SCI (22.6%),11 and higher than the rate of smoking in the general population in South Carolina (22.5%).24 The rate observed here also far exceeds other reports of persons with disability in the US (22.7%).25 As with previous research,11 we found smoking rates to be higher among those with less severe injuries. This may be due to increased physical disability among those with severe SCI which could limit independent smoking. Our cohort had a larger proportion of less severe injuries than most SCI cohorts owing to participants being identified through a surveillance system and not through a rehabilitation setting. However, smoking rates among the most severely injured in our study (25.5%) were still higher than in previous research (16.9%).11

This study was conducted in a state with a high rural population. Previous research has linked smoking with poor access to care in rural communities.16 Additionally, persons with disability are significantly more likely to live in poverty.26 In this cohort, 60.2% had a total household income of less than $25 000 per year, which is higher than previous reports within populations with SCI.23 In the general population, there are differences in smoking rates by income, with persons living below poverty level being more likely to be current smokers than those at or above poverty level.25 Additionally, when income increases above poverty level, persons are less likely to continue smoking.27 We also found poorer access to care related to a greater likelihood of being a current smoker, no association to trying to quit smoking, but decreased likelihood of using cessation support among those who did attempt to quit. Research in the general population shows persons with lower SES and/or without health insurance are less likely to receive smoking advice from a physician;15 therefore, those with poor access to care may want to quit but lack the essential resources and access to professionals that can enhance the likelihood of success.

In our study, of the current smokers who had ever tried to quit, 55.5% had made an attempt in the past year, and of these, 27.9% had ever sought professional help. The proportion reporting ever seeking professional help is higher than one study of persons with mobility impairments (5.3%), but that study only included counseling.28 Of those in the general population who made an attempt in the past year, 36.1% had used behavioral or pharmacological treatments.29 Similar to our results, Borrelli et al.28 found that 57.3% attempted to quit in the past year, and in the general population, 52% of smokers make a quit attempt each year.25 These findings indicate a large portion of people would like to quit smoking and suggest the need for identifying effective, accessible interventions that account for the specific needs of those with mobility impairment related to SCI such as physical limitations and inadequate access to care. We found different variables to be associated with quit attempts (younger age, lack of insurance) compared with use of cessation assistance (female gender, white race, older age and having insurance). These results are partially supported in the general population, with those in younger age groups being more likely to have a quit attempt, but quit attempts did not differ by health plan type.30 The same study also found women, those with older age, and having insurance to be more likely to receive professional advice to quit.

Limitations

Although this study provides valuable information on smoking in a population-based cohort of persons with SCI, there are limitations to be noted. First, the present study relied solely on self-reported data, without corroboration via medical records or other independent sources. However, we limited bias by relying on established population surveillance methods (BRFSS). Second, this study was cross-sectional in nature, and thus the nature of the data collection methods precluded any assessment of the identified associations over time (that is, longitudinally). Third, our response rate was 40.4%; therefore, we are not able to ascertain smoking status on those who did not respond. Lastly, we did not have information concerning military history as participants were recruited from a civilian rehabilitation hospital. Thus, findings may not be generalizable to a noncivilian population.

Conclusion

Rates of smoking within SCI populations far exceed national norms, but rates for making quit attempts are generally the same. Smokers with SCI face unique challenges to cessation, and thus future studies should examine the scope, reach and efficacy of cessation interventions, both pharmacologic and behavioral. Persons with SCI have physical impairments that may affect smoking cessation in various ways, such as the need for assistance from other persons or use of assistive devices. Additionally, professional help and medication for cessation may be inaccessible to those with SCI, especially those in rural areas. These considerations should be taken into account when creating interventions for smoking cessation in persons with SCI.

Data archiving

There were no data to deposit.