Abstract
Objectives. To examine associations between patient factors and smoking cessation assistance in US safety-net clinics.
Methods. Using electronic health record data from the OCHIN network, we identified adults with at least 1 primary care visit to a study clinic (n = 143 clinics in 12 states) with at least 1 documented “current smoker” status during 2014 to 2016 (n = 136 314; 29.8%). We estimated odds ratios (ORs) of smoking cessation assistance receipt (none [reference], counseling, medication, or both) by patient covariates.
Results. For all cessation assistance categories, odds of assistance were higher among women, those with more visits, those assessed and ready to quit, and patients with asthma or chronic obstructive pulmonary disease and hyperlipidemia. Odds of receiving both counseling and medication were lower among uninsured patients (OR = 0.56; 95% confidence interval [CI] = 0.48, 0.64), those of a race/ethnicity other than non-Hispanic White (OR range = 0.65–0.82), and those with diabetes (OR = 0.85; 95% CI = 0.79, 0.92), and higher among older patients and those with a comorbidity, with few exceptions.
Conclusions. Disparities in smoking cessation assistance receipt exist in safety-net settings, in particular by health insurance coverage and across race/ethnicity, even after control for other socioeconomic and demographic factors.
The United States has seen a marked decrease in smoking rates among adults in the past decade, with current smoking at 15.5% in 2016.1 Despite these reductions among the general population, significant disparities in smoking rates persist and the need to address these inequalities has become a focus of public health agendas.2,3 Substantially higher than national rates are seen among adults who live below the poverty line (25.3%), Medicaid recipients (25.3%), uninsured adults (28.4%), and individuals with serious psychological distress (35.8%).1
Strategies to reduce tobacco-related disparities include improving the availability, accessibility, and effectiveness of smoking cessation services for populations with high smoking prevalence. The recently updated National Commission on Prevention Priorities list continues to rank tobacco use screening and assistance among the 3 highest-priority preventive services and one of the few that are cost-saving.4 Evidence-based smoking cessation treatments in primary care settings constitute standard care5; however, rates of provision of appropriate cessation services by health care providers remain highly variable.6–8 Two initiatives that could have an impact on treatment among vulnerable populations are the 2014 Affordable Care Act’s (ACA’s) mandate for smoking cessation assistance for patients with certain insurance coverage,9,10 and implementation of the Meaningful Use of Electronic Health Records (EHR) initiative,11 which includes smoking cessation assistance measures.
Safety net settings, such as community health centers (CHCs) and public health clinics, provide health care to uninsured, Medicaid, and other vulnerable patients.12 These settings are important resources for reducing smoking disparities through cessation assistance, as they serve patients with a high prevalence of smoking.13–15 Although research has documented racial, ethnic, and socioeconomic disparities in the receipt of smoking cessation advice and aids, these studies have primarily been conducted with data from patients of more traditional primary care offices and often rely on self-reported surveys.16–22 Research has suggested that disparities in smoking rates are attributable, in part, to lack of access to care and appropriate treatment of smoking cessation3; these differences could be mitigated in safety clinics that provide increased access to care.23
Few studies have examined whether the characteristics found to be correlated with cessation assistance in other settings, including age, race/ethnicity, insurance status, and household income, are also associated with smoking-cessation assistance in CHC settings.13,14 Furthermore, none have utilized objective EHR data from a network of safety-net clinics across the United States after implementation of policy initiatives such as the ACA and Meaningful Use. The current study aims to fill in this gap by examining the patient factors associated with smoking cessation assistance (smoking cessation medication orders or counseling or both) among 136 314 patients identified as smokers from a network of 143 CHCs, nonprofit clinics, and county clinics across 12 states. It is hypothesized that there will be fewer disparities in smoking cessation assistance by patient characteristics among this population because of both decreased barriers to care and new policies focused on increasing smoking cessation assistance in primary care settings.
METHODS
OCHIN (not an acronym) is a nonprofit health information technology organization that provides a single, linked (each patient has a single identification number and medical record shared across every clinic in the network) instance of the Epic EHR.24 We extracted data from 143 safety net clinics (1 county clinic, 121 federally qualified health centers, 19 nonprofit clinics, and 2 rural health clinics) on the OCHIN network that were using OCHIN’s electronic health records by January 1, 2013, from structured EHR fields.
Study Population
We included adults aged 18 years and older as of January 1, 2014, with at least 1 primary care visit to a study clinic and identified as a “current smoker” at any visit during 2014 through 2016. We linked individual patient data across the 3 years and categorized a patient as either a “current smoker” or “not a current smoker”; that is, if a patient was identified as a smoker in 2015 but identified as a former smoker in 2016, they were included in the study population as a “current smoker” because provision of cessation services was possible. We excluded patients who were pregnant at any point during the study period (n = 6732 pregnant patients with indication of current smoking status), as the recommendations for smoking cessation assistance differ for pregnant women.
Variables
Current smoking status.
The denominator for smoking cessation assistance included patients who met the study population criteria and were identified as a current smoker during at least 1 visit in 2014 through 2016. A patient was identified as a current smoker if smoking status in the discrete data field in the vital signs or social history was “current every day smoker”; “current some day smoker”; “smoker, current status unknown”; “heavy tobacco smoker”; or “light tobacco smoker.” Changes made within the EHR are date stamped and saved; thus, a patient was considered to be a current smoker if the date stamped in the discrete field at any visit was between 2014 and 2016.
Smoking cessation assistance.
We dichotomized all smoking assistance outcomes (yes vs no); “yes” represents that the assistance occurred during at least 1 visit during the study period. We extracted smoking cessation medication orders from the medication orders list for the following medications: bupropion, varenicline, and all nicotine replacement therapy products.
Receipt of counseling was deemed “yes” if the discrete field “counseling given” was coded as “yes,” or by standard procedure codes for smoking cessation counseling or an internal OCHIN Epic code for counseling referral.
If a patient had both a smoking cessation medication ordered and documented receipt of counseling at any point in the study period, that patient was considered to have received both treatments, regardless of whether they occurred on the same office visit.
Patient characteristics.
We extracted discrete data from the EHR for the following characteristics known to be associated with smoking rates or smoking cessation assistance in primary care13,14,18,25,26: gender, age, race/ethnicity, household income as assessed through percentage of federal poverty level (FPL) according to the US Department of Health and Human Services, number of visits, insurance coverage at the majority of visits, and medical and psychiatric comorbidities as identified via the International Classification of Diseases, Ninth and Tenth Revisions (Geneva, Switzerland: World Health Organization; 1980 and 1992) codes (hypertension, diabetes, hyperlipidemia, asthma or chronic obstructive pulmonary disease [COPD], coronary artery disease [CAD], cancer, psychiatric disorder, substance use disorder other than tobacco use disorder). We also included readiness to quit smoking, a discrete field in the EHR (accessible via the vital signs and social history tabs) with 3 possible values: not populated (not assessed), checked “yes” (assessed and ready to quit), or checked “no” (assessed and not ready to quit). We also identified whether a patient’s clinic was urban or rural, identified with rural–urban commuting area codes.
Analyses
We first described the patient characteristics of those identified as current smokers compared with those not identified as a current smoker during the study period. We compared patient characteristics of smokers and nonsmokers by using absolute standardized differences, which are statistical measures of effect size between groups that are not influenced by large sample sizes. We considered a standardized difference of greater than 0.1 to denote marginal differences between the groups.27 We then examined characteristics of the current smoker cohort, overall and stratified by receipt of smoking cessation assistance. Next, we fit a multivariable multinomial logistic regression model to assess the association between patient characteristics and the receipt of smoking cessation assistance type (4 categories: none [serving as the reference], counseling received, medication order, or both). We also included indicator variables for the service area (independent health care organizations that have 1 or more clinics and often are in geographic proximity to one another) and the state the patient’s primary clinic is located in to adjust for clustering of patients within service area and to account for potential differences between service areas and states.
We conducted all statistical analyses in 2017 with SAS Enterprise Guide, version 7.13 (SAS Institute Inc, Cary, NC). All statistical tests were 2-sided and we defined significance as P < .05.
RESULTS
Of the 457 344 nonpregnant patients, 98.6% had smoking status documented at least once. Of those, 136 314 (29.8%) were identified as a current smoker at some point during this time (Table 1). Current smokers differed from nonsmokers on most characteristics. The current smoker sample was primarily non-Hispanic White (66.7%), had household incomes less than or equal to 138% of the FPL, and was insured by Medicaid (63.6%). Comorbidities ranged from a low of 2.3% for CAD to 44.5% for psychiatric disorder.
TABLE 1—
Characteristics of Study Patients by Smoking Status: OCHIN Network, United States, 2014 to 2016
| Characteristic | Current Smoker, No. (%) | Not a Current Smoker, No. (%) | Absolute Standardized Differencea |
| Total | 136 314 (29.8) | 321 030 (70.2) | |
| Gender | 0.22 | ||
| Male | 66 042 (48.4) | 120 167 (37.4) | |
| Female | 70 272 (51.6) | 200 863 (62.6) | |
| Age,b y | 0.32 | ||
| 18–24 | 14 985 (11.0) | 42 791 (13.3) | |
| 25–34 | 31 589 (23.2) | 68 113 (21.2) | |
| 35–44 | 27 472 (20.2) | 60 566 (18.9) | |
| 45–54 | 32 412 (23.8) | 55 591 (17.3) | |
| 55–64 | 23 283 (17.1) | 52 446 (16.3) | |
| ≥ 65 | 6 573 (4.8) | 41 523 (12.9) | |
| Race/ethnicity | 0.57 | ||
| Non-Hispanic White | 90 891 (66.7) | 157 673 (49.1) | |
| Hispanic | 13 238 (9.7) | 99 340 (30.9) | |
| Non-Hispanic Black | 24 493 (18.0) | 41 270 (12.9) | |
| Non-Hispanic other | 4 874 (3.6) | 15 285 (4.8) | |
| Unknown | 2 818 (2.1) | 7 462 (2.3) | |
| Federal poverty levelc | 0.21 | ||
| ≤ 138% | 99 318 (72.9) | 202 607 (63.1) | |
| > 138% | 20 658 (15.2) | 69 781 (21.7) | |
| Missing | 16 338 (12.0) | 48 642 (15.2) | |
| Location of primary clinic | 0.05 | ||
| Urban | 122 197 (89.6) | 282 526 (88.0) | |
| Rural | 13 006 (9.5) | 35 482 (11.1) | |
| Unknown | 1 111 (0.8) | 3 022 (0.9) | |
| Number of visits | 0.04 | ||
| 1 | 40 686 (29.8) | 99 406 (31.0) | |
| 2–5 | 61 605 (45.2) | 146 958 (45.8) | |
| ≥ 6 | 34 023 (25.0) | 74 666 (23.3) | |
| Insurance coverage atmajority of visits | 0.43 | ||
| Commercial | 15 713 (11.5) | 69 845 (21.8) | |
| Medicaid | 86 647 (63.6) | 140 708 (43.8) | |
| Medicare | 17 054 (12.5) | 47 247 (14.7) | |
| Self-pay (uninsured) | 16 139 (11.8) | 59 157 (18.4) | |
| Other | 761 (0.6) | 4 073 (1.3) | |
| Readiness to quit | 1.42 | ||
| Ready to quit assessed, not ready to quit | 35 054 (25.7) | 3 154 (1.0) | |
| Ready to quit not assessed | 64 208 (47.1) | 316 582 (98.6) | |
| Ready to quit assessed, ready to quit | 37 052 (27.2) | 1 294 (0.4) | |
| Smoking status assessed in study period | 0.19 | ||
| No | 0 (0.0) | 5 794 (1.8) | |
| Yes | 136 314 (100.0) | 315 236 (98.2) | |
| Comorbidities | |||
| Hypertension | 38 696 (28.4) | 91 874 (28.6) | 0.01 |
| Diabetes | 16 274 (11.9) | 49 252 (15.3) | 0.10 |
| Asthma or COPD | 27 123 (19.9) | 35 773 (11.1) | 0.24 |
| Coronary artery disease | 3 141 (2.3) | 7 358 (2.3) | 0.01 |
| Hyperlipidemia | 31 584 (23.2) | 84 604 (26.4) | 0.07 |
| Cancer | 7 624 (5.6) | 19 982 (6.2) | 0.03 |
| Psychiatric disorderd | 60 626 (44.5) | 87 293 (27.2) | 0.37 |
| Substance use disordere | 39 621 (29.1) | 20 568 (6.4) | 0.62 |
Note. COPD = chronic obstructive pulmonary disease. Patients were included if they had at least 1 visit to a study clinic in 2014 through 2016 and were not pregnant at any point during the study period. P values comparing demographics between smokers and not current smokers using the χ2 test were all P < .001, except for hypertension (P = .113) and coronary artery disease (P = .800).
Absolute standardized difference greater than 0.1 indicates significant differences between the 2 groups.
Age as of January 1, 2014.
Federal poverty level is estimated according to the US Department of Health and Human Services.
Psychiatric disorders included anxiety disorders and posttraumatic stress disorder, schizophrenia and other psychosis disorders, depressive disorders, and bipolar disorder.
Substance use disorder excluded tobacco use disorder.
Current Smokers and Smoking Cessation Assistance Type
Table 2 shows the characteristics of the cohort of patients identified as current smokers by smoking cessation assistance types. Overall, 46.2% received no smoking cessation assistance during the study period, 35.2% received counseling only, 7.5% received a smoking cessation medication order, and 11.1% received the recommended treatment of both medication and counseling. Distributions varied by patient factors across smoking cessation assistance.
TABLE 2—
Characteristics of Current Smokers by Smoking Cessation Assistance Type: OCHIN Network, United States, 2014 to 2016
| Smoking Cessation Assistance Type |
||||
| Characteristics | No Assistance, No. (%) | Counseling Only,a No. (%) | Medication Ordered Only,b No. (%) | Both Counseling and Medication Ordered, No. (%) |
| Total | 63 004 (46.2) | 47 976 (35.2) | 10 179 (7.5) | 15 155 (11.1) |
| Gender | ||||
| Male | 32 175 (51.1) | 23 009 (48.0) | 4 588 (45.1) | 6 270 (41.4) |
| Female | 30 829 (48.9) | 24 967 (52.0) | 5 591 (54.9) | 8 885 (58.6) |
| Age,c y | ||||
| 18–24 | 8 412 (13.4) | 5 275 (11.0) | 595 (5.8) | 703 (4.6) |
| 25–34 | 16 419 (26.1) | 10 810 (22.5) | 1 933 (19.0) | 2 427 (16.0) |
| 35–44 | 12 831 (20.4) | 9 482 (19.8) | 2 145 (21.1) | 3 014 (19.9) |
| 45–54 | 13 298 (21.1) | 11 449 (23.9) | 2 924 (28.7) | 4 741 (31.3) |
| 55–64 | 9 079 (14.4) | 8 484 (17.7) | 2 122 (20.8) | 3 598 (23.7) |
| ≥ 65 | 2 965 (4.7) | 2 476 (5.2) | 460 (4.5) | 672 (4.4) |
| Race/ethnicity | ||||
| Non-Hispanic White | 42 194 (67.0) | 30 018 (62.6) | 7 906 (77.7) | 10 773 (71.1) |
| Hispanic | 7 117 (11.3) | 4 666 (9.7) | 617 (6.1) | 838 (5.5) |
| Non-Hispanic Black | 9 875 (15.7) | 10 667 (22.2) | 1 141 (11.2) | 2 810 (18.5) |
| Non-Hispanic other | 2 384 (3.8) | 1 660 (3.5) | 365 (3.6) | 465 (3.1) |
| Unknown | 1 434 (2.3) | 965 (2.0) | 150 (1.5) | 269 (1.8) |
| Federal poverty leveld | ||||
| ≤ 138% | 43 390 (68.9) | 36 651 (76.4) | 7 383 (72.5) | 11 894 (78.5) |
| > 138% | 9 765 (15.5) | 7 065 (14.7) | 1 647 (16.2) | 2 181 (14.4) |
| Unknown | 9 849 (15.6) | 4 260 (8.9) | 1 149 (11.3) | 1 080 (7.1) |
| Location of primary clinic | ||||
| Urban | 56 303 (89.4) | 42 979 (89.6) | 9 175 (90.1) | 13 740 (90.7) |
| Rural | 6 206 (9.9) | 4 612 (9.6) | 910 (8.9) | 1 278 (8.4) |
| Unknown | 495 (0.8) | 385 (0.8) | 94 (0.9) | 137 (0.9) |
| Number of visits | ||||
| 1 | 26 377 (41.9) | 11 411 (23.8) | 1 798 (17.7) | 1 100 (7.3) |
| 2–5 | 27 975 (44.4) | 23 104 (48.2) | 4 708 (46.3) | 5 818 (38.4) |
| ≥ 6 | 8 652 (13.7) | 13 461 (28.1) | 3 673 (36.1) | 8 237 (54.4) |
| Insurance coverage at majority of visits | ||||
| Commercial | 8 104 (12.9) | 5 458 (11.4) | 945 (9.3) | 1 206 (8.0) |
| Medicaid | 37 849 (60.1) | 30 448 (63.5) | 7 202 (70.8) | 11 148 (73.6) |
| Medicare | 6 997 (11.1) | 6 260 (13.0) | 1 482 (14.6) | 2 315 (15.3) |
| Self-pay (uninsured) | 9 585 (15.2) | 5 548 (11.6) | 530 (5.2) | 476 (3.1) |
| Other | 469 (0.7) | 262 (0.5) | 20 (0.2) | 10 (0.1) |
| Readiness to quit smoking | ||||
| Not assessed | 52 408 (83.2) | 3 790 (7.9) | 7 389 (72.6) | 621 (4.1) |
| Assessed, not ready to quit | 6 446 (10.2) | 23 881 (49.8) | 982 (9.6) | 3 745 (24.7) |
| Assessed, ready to quit | 4 150 (6.6) | 20 305 (42.3) | 1 808 (17.8) | 10 789 (71.2) |
| Comorbidities | ||||
| Hypertension | 14 264 (22.6) | 14 804 (30.9) | 3 325 (32.7) | 6 303 (41.6) |
| Diabetes | 5 839 (9.3) | 6 478 (13.5) | 1 291 (12.7) | 2 666 (17.6) |
| Asthma or COPD | 9 685 (15.4) | 9 162 (19.1) | 3 008 (29.6) | 5 268 (34.8) |
| Coronary artery disease | 1 071 (1.7) | 1 083 (2.3) | 332 (3.3) | 655 (4.3) |
| Hyperlipidemia | 11 285 (17.9) | 11 861 (24.7) | 2 954 (29.0) | 5 484 (36.2) |
| Cancer | 2 947 (4.7) | 2 594 (5.4) | 811 (8.0) | 1 272 (8.4) |
| Psychiatric disordere | 25 165 (39.9) | 21 267 (44.3) | 5 557 (54.6) | 8 637 (57.0) |
| Substance use disorderf | 15 513 (24.6) | 14 491 (30.2) | 3 815 (37.5) | 5 802 (38.3) |
Note. COPD = chronic obstructive pulmonary disease.
Counseling included “counseling given” in the electronic health records vital sign checked, standard procedure codes for smoking counseling assistance, and OCHIN internal codes for referral for services.
Smoking cessation medication included bupropion, varenicline, and nicotine replacement products.
Age as of January 1, 2014.
Federal poverty level is estimated according to the US Department of Health and Human Services.
Psychiatric disorders included anxiety disorders and posttraumatic stress disorder, schizophrenia and other psychosis disorders, depressive disorders, and bipolar disorder.
Substance use disorder excluded tobacco use disorder.
Smoking Cessation Counseling vs No Assistance
Female gender, higher numbers of visits, being assessed and ready to quit, hypertension, asthma or COPD, and hyperlipidemia were associated with higher odds of counseling versus no assistance (Table 3). Odds of counseling receipt did not differ by FPL or by diagnoses of diabetes, CAD, cancer, psychiatric disorder, or substance use disorder. There were no significant differences in the odds of receipt of counseling by rural versus urban location of primary care clinic or by race/ethnicity except between non-Hispanic other race and non-Hispanic White. Medicare insurance was associated with lower odds of counseling than those with commercial insurance and patients aged 18 to 24 years had higher odds of receipt of counseling than did those in the age categories of 25 to 34, 35 to 44, and 45 to 54 years; odds did not differ significantly between the reference group and the oldest age categories. Finally, those not assessed for readiness to quit had significantly lower odds of receipt of counseling compared with those assessed and not ready to quit.
TABLE 3—
Adjusted Odds Ratios of Smoking Cessation Assistance Type Versus No Assistance by Patient Characteristics Among Current Smokers: OCHIN Network, United States, 2014 to 2016
| Smoking Cessation Assistance Type (Ref = No Assistance) |
|||
| Characteristic | Counseling Only,a,b
OR (95% CI) |
Medication Ordered Only,a,c OR (95% CI) | Counseling and Medication Ordered,a,b,c OR (95% CI) |
| Gender | |||
| Male (Ref) | 1 | 1 | 1 |
| Female | 1.08 (1.04, 1.12) | 1.07 (1.02, 1.12) | 1.18 (1.13, 1.24) |
| Age,d y | |||
| 18–24 (Ref) | 1 | 1 | 1 |
| 25–34 | 0.93 (0.87, 0.99) | 1.50 (1.36, 1.65) | 1.39 (1.25, 1.55) |
| 35–44 | 0.92 (0.86, 0.99) | 1.92 (1.74, 2.12) | 1.77 (1.59, 1.97) |
| 45–54 | 0.92 (0.86, 0.98) | 2.20 (1.99, 2.43) | 1.94 (1.75, 2.16) |
| 55–64 | 0.98 (0.91, 1.06) | 2.13 (1.92, 2.37) | 1.92 (1.72, 2.15) |
| ≥ 65 | 0.96 (0.85, 1.08) | 1.39 (1.20, 1.62) | 1.24 (1.05, 1.46) |
| Race/ethnicity | |||
| Non-Hispanic White (Ref) | 1 | 1 | 1 |
| Hispanic | 1.00 (0.93, 1.07) | 0.61 (0.55, 0.67) | 0.65 (0.58, 0.71) |
| Non-Hispanic Black | 0.97 (0.91, 1.04) | 0.80 (0.73, 0.87) | 0.82 (0.75, 0.89) |
| Non-Hispanic other | 0.86 (0.78, 0.95) | 0.86 (0.77, 0.97) | 0.74 (0.65, 0.84) |
| Unknown | 1.02 (0.90, 1.16) | 0.78 (0.66, 0.93) | 0.98 (0.82, 1.17) |
| Federal poverty levele | |||
| ≤ 138% (Ref) | 1 | 1 | 1 |
| > 138% | 1.03 (0.98, 1.09) | 1.01 (0.95, 1.08) | 1.05 (0.98, 1.13) |
| Unknown | 0.99 (0.92, 1.06) | 0.94 (0.86, 1.04) | 0.96 (0.86, 1.07) |
| Location of primary clinic | |||
| Urban (Ref) | 1 | 1 | 1 |
| Rural | 1.05 (0.97, 1.14) | 0.96 (0.87, 1.06) | 0.98 (0.88, 1.10) |
| Unknown | 0.80 (0.66, 0.98) | 0.94 (0.75, 1.19) | 0.75 (0.58, 0.97) |
| Number of visits | |||
| 1 (Ref) | 1 | 1 | 1 |
| 2–5 | 1.69 (1.61, 1.76) | 1.90 (1.79, 2.01) | 3.15 (2.91, 3.40) |
| ≥ 6 | 2.71 (2.56, 2.87) | 3.60 (3.36, 3.87) | 9.17 (8.41, 10.00) |
| Insurance coverage at majority of visits | |||
| Commercial (Ref) | 1 | 1 | 1 |
| Medicaid | 0.96 (0.90, 1.02) | 1.33 (1.22, 1.44) | 1.17 (1.07, 1.28) |
| Medicare | 0.90 (0.83, 0.98) | 1.13 (1.02, 1.25) | 0.93 (0.84, 1.04) |
| Self-pay (uninsured) | 0.96 (0.88, 1.04) | 0.58 (0.51, 0.67) | 0.56 (0.48, 0.64) |
| Other | 0.92 (0.72, 1.18) | 0.58 (0.37, 0.92) | 0.25 (0.13, 0.49) |
| Readiness to quit smoking | |||
| Assessed and not ready to quit (Ref) | 1 | 1 | 1 |
| Assessed and ready to quit | 1.23 (1.17, 1.29) | 2.67 (2.44, 2.91) | 3.84 (3.62, 4.07) |
| Not assessed | 0.02 (0.016, 0.018) | 1.10 (1.02, 1.19) | 0.02 (0.02, 0.03) |
| Comorbiditiesf | |||
| Hypertension | 1.08 (1.03, 1.14) | 1.04 (0.98, 1.10) | 1.18 (1.11, 1.25) |
| Diabetes | 0.99 (0.93, 1.05) | 0.85 (0.79, 0.91) | 0.85 (0.79, 0.92) |
| Asthma or COPD | 1.07 (1.02, 1.12) | 1.60 (1.52, 1.69) | 1.70 (1.60, 1.80) |
| Coronary artery disease | 1.12 (0.99, 1.28) | 1.24 (1.08, 1.42) | 1.53 (1.32, 1.78) |
| Hyperlipidemia | 1.20 (1.14, 1.26) | 1.21 (1.14, 1.28) | 1.41 (1.32, 1.49) |
| Cancer | 0.96 (0.88, 1.04) | 1.16 (1.06, 1.26) | 1.07 (0.97, 1.18) |
| Psychiatric disorderg | 0.99 (0.95, 1.03) | 1.09 (1.04, 1.14) | 1.08 (1.03, 1.14) |
| Substance use disorderh | 1.04 (0.997, 1.08) | 1.31 (1.24, 1.37) | 1.24 (1.18, 1.31) |
Note. CI = confidence interval; COPD = chronic obstructive pulmonary disease; OR = odds ratio. n = 136 314. Adjusted odds ratios were estimated by using a multivariable multinomial logistic regression model with service areas and states included as fixed effects through a set of indicator variables.
Compared with no documented smoking cessation assistance during the study period.
Counseling included “counseling given” in the electronic health records vital sign checked, standard procedure codes for smoking counseling assistance, and OCHIN internal codes for referral for services.
Smoking cessation medication included bupropion, varenicline, and nicotine replacement products.
Age as of January 1, 2014.
Federal poverty level is estimated according to the US Department of Health and Human Services.
Reference for each comorbid condition is the absence of the condition.
Psychiatric disorders included anxiety disorders and posttraumatic stress disorder, schizophrenia and other psychosis disorders, depressive disorders, and bipolar disorder.
Substance use disorder excluded tobacco use disorder.
Cessation Medication Order vs No Assistance
There were no significant differences in the odds of having a cessation medication ordered versus no assistance by FPL, rural location, or diagnosis of hypertension. Older age was associated with higher odds of medication orders, with a range of ORs of 1.50 for those aged 25 to 34 years to 2.20 for those aged 45 to 54 years, compared with patients aged 18 to 24 years. Women had 7% increased odds of having an order compared with men, and non-Hispanic Whites had significantly higher odds compared with all other races/ethnicities. Having at least 6 visits compared with just 1 visit in the study period was associated with greater than 3-fold higher odds of having a medication order. Patients with public insurance had higher odds of having a medication ordered compared with commercially insured patients, but being uninsured was associated with lower odds (OR = 0.58; 95% CI = 0.51, 0.67). Compared with patients assessed and not ready to quit, having readiness assessed and being ready to quit was associated with more than 2-times-higher odds of having a medication order, and patients who were not assessed had 10% increased odds. Finally, having a diagnosis of asthma or COPD, CAD, hyperlipidemia, cancer, psychiatric disorder, or substance use disorder resulted in higher odds of medication ordered versus no assistance, whereas a diagnosis of diabetes was associated with lower odds.
Medication and Counseling vs No Assistance
Predictors of having both a medication ordered and counseling receipt versus no assistance were similar to those of medication order only: higher odds among all older age categories compared with patients aged 18 to 24 years, those with more visits (≥ 6 visits vs 1; OR = 9.17; 95% CI = 8.41, 10.00), women compared with men, and those with asthma or COPD, CAD, hyperlipidemia, psychiatric disorder, and substance use disorder. Similar to medication orders only, odds of receipt of both medications and counseling was lower among all races/ethnicities than among non-Hispanic Whites (except unknown category) and patients with diabetes; there were no differences in odds by percentage of FPL or by urban versus rural clinic location. Compared with commercially insured patients, those with Medicaid insurance had 17% higher odds of comprehensive smoking cessation assistance, and uninsured had lower odds; Medicare was not associated with odds of assistance. Finally, similar to counseling given, odds of receipt of both types of assistance were higher among those with hypertension and those assessed and ready to quit, lower among those not assessed for readiness to quit, and not significantly associated with a cancer diagnosis.
DISCUSSION
This is the first study to examine the predictors of smoking cessation assistance using EHR data from a large number of smokers seen in settings that provide health care to traditionally underserved populations. The high rates of smoking status assessment (98.7%) allowed us to examine smoking cessation assistance among a large, comprehensive cohort of smokers from safety-net clinics across the country.
Our hypothesis that there would be fewer disparities in smoking cessation assistance in safety-net clinics because of increased access to care among vulnerable populations was partially supported. Most notably, household income as assessed by FPL was not associated with smoking cessation assistance, a commonly reported characteristic associated with tobacco use disorder health disparities.3,21 A study using 2009 national survey data from adult community health center patients also reported that FPL was not associated with receipt of tobacco counseling. These results suggest that safety-net clinics are providing smoking cessation assistance equally to patients with varying levels of socioeconomic status. Furthermore, patients with Medicaid insurance had similar odds of counseling only, and higher odds of having a medication ordered and both medication and counseling, compared with those with commercial insurance. Survey-based studies conducted after ACA expansion with data from patients in non–safety net settings have shown similar results.21,28 This suggests that primary care providers are assisting Medicaid recipients, a population with disparately high rates of smoking, in their cessation efforts at the same rate as their commercially insured counterparts. As almost 90% of our study population were from Medicaid expansion states, we were not able to examine if there were differences in receipt of cessation assistance among states that expanded Medicaid via the ACA versus those that did not expand. This is an important research question that could have significant policy implications.
Our hypothesis was partially refuted by the finding that disparities by race/ethnicity and uninsured status were similar to those of other studies among patients in general primary care settings3,18–21,29–31 for medication orders and the more comprehensive assistance of both medication and counseling. Interestingly, this study did not find these same disparities for counseling only. Within the CHCs in our study, “counseling given” is a discrete button in the EHR that is easily clicked by providers at most or all relevant encounters and the patient is not necessarily billed for this procedure; thus, counseling might be more universally provided to those without insurance, as opposed to pharmacotherapy, which would likely be paid out of pocket. One explanation for the difference in provision of pharmacotherapy, but not counseling, among Hispanic and non-Hispanic Black patients is that they are more likely to be light and intermittent smokers.32,33 Providers might be less likely to prescribe pharmacotherapy to patients who are not heavy or daily users. It is also possible that pharmacotherapy for cessation is less acceptable or there are other barriers to availability among these patients. Despite lower levels of smoking and similar desires to quit smoking,3 successful cessation is substantially lower among these groups than among non-Hispanic Whites.3 Future research is needed to further our understanding of the etiology of these disparities.
Similar to other studies, we found that, with few exceptions, women, older patients (except for counseling only), those ready to quit, those with more visits, and patients with comorbidities had higher odds of receipt of cessation assistance.13,14,21,28,31 The youngest age group had lower odds of having a medication order only and both counseling and medication provision compared with all age groups. This younger cohort likely has a shorter smoking history and tends to be healthier; thus, providers might be less likely to order pharmacotherapy for these patients. Continued efforts to promote evidence-based cessation assistance, including pharmacotherapy, among young adults is vital to reducing the long-term negative effects of smoking.
Diabetes was the only comorbidity associated with lower odds of medication orders and both medication and counseling. Among patients with diabetes, it may be that providers are hesitant to provide smoking cessation medications because of the limited knowledge of the potential or unknown side effects of cessation medications among this patient population, the possible short-term worsening of some diabetic symptoms, or patient concerns about modest weight gain following smoking cessation34,35; further research is warranted. Not surprisingly, patients with asthma or COPD had the highest odds of most types of assistance, compared with the other comorbidities. However, the odds of receipt of both counseling and medication by comorbidity ranged from 0.85 for patients with diabetes to 1.70 for those with asthma or COPD. Many studies have examined the association of smoking cessation assistance with comorbidities but often do include these diseases individually. Our finding that the significance of the differences in odds varied by comorbidity and by type of assistance suggests that future studies should look at individual diseases and assistance type to fully understand these associations. Furthermore, although providing smoking cessation assistance to those with comorbidities is important as secondary and tertiary prevention, providers should ensure that smokers without medical conditions also receive assistance as primary prevention to avoid development of smoking-related diseases.
Limitations
This study had some limitations. We were able to assess whether medications were ordered, but not whether the patients filled the prescriptions and took the medications. We likely did not capture all patients who were using nicotine replacement therapy, as the 2 most common (patch and gum) are over the counter; thus, these might have been prescribed or discussed but not ordered via the EHR. This might be more true for uninsured patients who would not need a prescription to be reimbursed by insurance but might be more likely to be referred to a quit line and provided nicotine replacement therapy. We also were unable to assess if bupropion was prescribed for smoking cessation or depression; however, all patients in the study sample were current smokers and our models controlled for depressive disorders. This study did not look at the impact of having multiple disparities associated with smoking cessation assistance; for example, it might be that Hispanic patients are less likely to have visits or to be insured, which could have synergistic effects on lower odds of assistance; this was beyond the scope of the current study and future research on these dynamics is warranted. Finally, although we did not find differences in assistance by whether a patient’s primary clinic was in an urban versus rural location, the majority of clinics in this study were in urban settings.
Public Health Implications
Although there were no disparities in assistance by patient socioeconomic status or urban versus rural location of primary care clinic, other patient characteristics, such as whether someone has insurance coverage and a patient’s race/ethnicity, remained highly predictive of prescribing smoking cessation pharmacotherapy and more comprehensive assistance of both medication orders and counseling provision. The finding of lower odds of comprehensive assistance among uninsured patients than among those with commercial insurance provides support for continued policy efforts to increase insurance coverage among our most vulnerable populations. Furthermore, given the high rates of smoking and the disproportionately low rates of cessation among patients with certain demographic characteristics, future mixed methods research is needed to determine the etiology of these differences to inform policies and interventions targeted at eliminating tobacco-related health disparities.
ACKNOWLEDGMENTS
This work was supported by funding from the National Institute on Drug Abuse, K23DA037453.
Preliminary results of this study were presented at the 2018 Society for Research on Nicotine & Tobacco 24th Annual Meeting in Baltimore, MD, February 22, 2018.
The authors gratefully acknowledge the OCHIN network and the OCHIN Practice-Based Research Network.
Note. The funding agency had no role in study design; collection, analysis, and interpretation of data; writing the article; and the decision to submit the article for publication. The research presented in this article is that of the authors and does not reflect the official policy of the National Institutes of Health.
HUMAN PARTICIPANT PROTECTION
This study was approved by the institutional review board at Oregon Health & Science University.
REFERENCES
- 1.Jamal A, Phillips E, Gentzke AS et al. Current cigarette smoking among adults—United States, 2016. MMWR Morb Mortal Wkly Rep. 2018;67(2):53–59. doi: 10.15585/mmwr.mm6702a1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Centers for Disease Control and Prevention. Best Practices User Guide: Health Equity in Tobacco Prevention and Control. Atlanta, GA: National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2015. [Google Scholar]
- 3.US National Cancer Institute. National Institutes of Health. Bethesda, MD: US Department of Health and Human Services; 2017. A Socioecological Approach to Addressing Tobacco-Related Health Disparities. National Cancer Institute Tobacco Control Monograph 22. NIH Publication No. 17-CA-8035A. [Google Scholar]
- 4.Maciosek MV, LaFrance AB, Dehmer SP et al. Updated priorities among effective clinical preventive services. Ann Fam Med. 2017;15(1):14–22. doi: 10.1370/afm.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fiore MC, Jaén CR, Baker TB . Treating Tobacco Use and Dependence: 2008 Update: Clinical Practice Guideline. Rockville, MD: US Department of Health and Human Services, Public Health Service; 2008. [Google Scholar]
- 6.Hazlehurst B, Sittig DF, Stevens VJ et al. Natural language processing in the electronic medical record: assessing clinician adherence to tobacco treatment guidelines. Am J Prev Med. 2005;29(5):434–439. doi: 10.1016/j.amepre.2005.08.007. [DOI] [PubMed] [Google Scholar]
- 7.Williams RJ, Masica AL, McBurnie MA et al. Documentation of the 5 As for smoking cessation by PCPs across distinct health systems. Am J Manag Care. 2014;20:e35–e42. [PMC free article] [PubMed] [Google Scholar]
- 8.Bartsch AL, Harter M, Niedrich J, Brutt AL, Buchholz A. A systematic literature review of self-reported smoking cessation counseling by primary care physicians. PLoS One. 2016;11(12):e0168482. doi: 10.1371/journal.pone.0168482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Department of Health and Human Services. Centers for Medicare & Medicaid Services. Medicaid program; eligibility changes under the Affordable Care Act of 2010. Final rule, Interim final rule. Fed Regist. 2012;77:17144–17217. [PubMed] [Google Scholar]
- 10.McAfee T, Babb S, McNabb S, Fiore MC. Helping smokers quit—opportunities created by the Affordable Care Act. N Engl J Med. 2015;372(1):5–7. doi: 10.1056/NEJMp1411437. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Department of Health and Human Services. Centers for Medicare and Medicaid Services. Medicare and Medicaid programs; electronic health record incentive program. Final rule. Fed Regist. 2010;75:44313–44588. [PubMed] [Google Scholar]
- 12.Institute of Medicine. America’s Health Care Safety Net: Intact But Endangered. Washington, DC: National Academies Press; 2000. [PubMed] [Google Scholar]
- 13.Lebrun-Harris LA, Fiore MC, Tomoyasu N, Ngo-Metzger Q. Cigarette smoking, desire to quit, and tobacco-related counseling among patients at adult health centers. Am J Public Health. 2015;105(1):180–188. doi: 10.2105/AJPH.2013.301691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Silfen SL, Cha J, Wang JJ, Land TG, Shih SC. Patient characteristics associated with smoking cessation interventions and quit attempt rates across 10 community health centers with electronic health records. Am J Public Health. 2015;105(10):2143–2149. doi: 10.2105/AJPH.2014.302444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Flocke SA, Hoffman R, Eberth JM et al. The prevalence of tobacco use at federally qualified health centers in the United States, 2013. Prev Chronic Dis. 2017;14:E29. doi: 10.5888/pcd14.160510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cokkinides VE, Halpern MT, Barbeau EM, Ward E, Thun MJ. Racial and ethnic disparities in smoking-cessation interventions: analysis of the 2005 National Health Interview Survey. Am J Prev Med. 2008;34(5):404–412. doi: 10.1016/j.amepre.2008.02.003. [DOI] [PubMed] [Google Scholar]
- 17.Solberg LI, Parker ED, Foldes SS, Walker PF. Disparities in tobacco cessation medication orders and fills among special populations. Nicotine Tob Res. 2010;12(2):144–151. doi: 10.1093/ntr/ntp187. [DOI] [PubMed] [Google Scholar]
- 18.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. doi: 10.1080/14622200701704905. [DOI] [PubMed] [Google Scholar]
- 19.Lopez-Quintero C, Crum RM, Neumark YD. Racial/ethnic disparities in report of physician-provided smoking cessation advice: analysis of the 2000 National Health Interview Survey. Am J Public Health. 2006;96(12):2235–2239. doi: 10.2105/AJPH.2005.071035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sonnenfeld N, Schappert SM, Lin SX. Racial and ethnic differences in delivery of tobacco-cessation services. Am J Prev Med. 2009;36(1):21–28. doi: 10.1016/j.amepre.2008.09.028. [DOI] [PubMed] [Google Scholar]
- 21.Tan ASL, Young-Wolff KC, Carter-Harris L, Salloum RG, Banerjee SC. Disparities in the receipt of tobacco treatment counseling within the US context of the Affordable Care Act and meaningful use implementation. Nicotine Tob Res. 2017 doi: 10.1093/ntr/ntx233. Epub ahead of print. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.King BA, Dube SR, Babb SD, McAfee TA. Patient-reported recall of smoking cessation interventions from a health professional. Prev Med. 2013;57(5):715–717. doi: 10.1016/j.ypmed.2013.07.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.National Association of Community Health Centers. A sketch of community health centers. Chart book 2017. Available at. http://www.nachc.org/wp-content/uploads/2017/06/Chartbook2017.pdf . Accessed April 10, 2018.
- 24.Devoe JE, Sears A. The OCHIN community information network: bringing together community health centers, information technology, and data to support a patient-centered medical village. J Am Board Fam Med. 2013;26(3):271–278. doi: 10.3122/jabfm.2013.03.120234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Japuntich SJ, Leventhal AM, Piper ME et al. Smoker characteristics and smoking-cessation milestones. Am J Prev Med. 2011;40(3):286–294. doi: 10.1016/j.amepre.2010.11.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Jamal A, King BA, Neff LJ, Whitmill J, Babb SD, Graffunder CM. Current cigarette smoking among adults—United States, 2005–2015. MMWR Morb Mortal Wkly Rep. 2016;65(44):1205–1211. doi: 10.15585/mmwr.mm6544a2. [DOI] [PubMed] [Google Scholar]
- 27.Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput. 2009;38(6):1228–1234. [Google Scholar]
- 28.Young-Wolff KC, Klebaner D, Campbell CI, Weisner C, Satre DD, Adams AS. Association of the Affordable Care Act with smoking and tobacco treatment utilization among adults newly enrolled in health care. Med Care. 2017;55(5):535–541. doi: 10.1097/MLR.0000000000000712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Papadakis S, Cole AG, Reid RD et al. Increasing rates of tobacco treatment delivery in primary care practice: evaluation of the Ottawa Model for Smoking Cessation. Ann Fam Med. 2016;14(3):235–243. doi: 10.1370/afm.1909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lopez MR, Jansen K, Souza LD, Pinheiro RT, Tomasi E, Silva RA. Prevalence and profile of daily smokers seen at three primary health care units in Pelotas, southern Brazil. Trends Psychiatry Psychother. 2012;34(3):154–160. doi: 10.1590/s2237-60892012000300007. [DOI] [PubMed] [Google Scholar]
- 31.Babb S, Malarcher A, Schauer G, Asman K, Jamal A. Quitting smoking among adults—United States, 2000–2015. MMWR Morb Mortal Wkly Rep. 2017;65(52):1457–1464. doi: 10.15585/mmwr.mm6552a1. [DOI] [PubMed] [Google Scholar]
- 32.Trinidad DR, Perez-Stable EJ, Emery SL, White MM, Grana RA, Messer KS. Intermittent and light daily smoking across racial/ethnic groups in the United States. Nicotine Tob Res. 2009;11(2):203–210. doi: 10.1093/ntr/ntn018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Trinidad DR, Perez-Stable EJ, White MM, Emery SL, Messer K. A nationwide analysis of US racial/ethnic disparities in smoking behaviors, smoking cessation, and cessation-related factors. Am J Public Health. 2011;101(4):699–706. doi: 10.2105/AJPH.2010.191668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Tonstad S. Cigarette smoking, smoking cessation, and diabetes. Diabetes Res Clin Pract. 2009;85(1):4–13. doi: 10.1016/j.diabres.2009.04.013. [DOI] [PubMed] [Google Scholar]
- 35.Nagrebetsky A, Brettell R, Roberts N, Farmer A. Smoking cessation in adults with diabetes: a systematic review and meta-analysis of data from randomised controlled trials. BMJ Open. 2014;4(3):e004107. doi: 10.1136/bmjopen-2013-004107. [DOI] [PMC free article] [PubMed] [Google Scholar]