Disparities in Access to Over-the-Counter Nicotine Replacement Products in New York City Pharmacies

Abstract

Objectives. We surveyed the availability of tobacco products and nonprescription nicotine replacement therapy (NRT) in pharmacies in New York City, stratified by the race, ethnicity, and socioeconomic status (SES) of the surrounding neighborhoods to determine whether disparities in availability existed.

Methods. Surveyors visited a random sample of retail pharmacies to record the availability of tobacco products and nonprescription NRT. We used census data and geographic information systems analysis to determine the SES of each neighborhood. We used logistic modeling to explore relations between SES and the availability of NRT and tobacco products.

Results. Of 646 pharmacies sampled, 90.8% sold NRT and 46.9% sold cigarettes. NRT and cigarettes were slightly more available in pharmacies in neighborhoods with a higher SES. NRT was more expensive in poorer neighborhoods.

Conclusions. Small disparities existed in access to nonprescription NRT and cigarettes. The model did not adequately account for cigarette access, because of availability from other retail outlets. These results may explain some of the excess prevalence of cigarette use in low-SES areas.

Smoking remains the leading cause of preventable death and disease in the United States, where 1 in 5, or 440 000, Americans die each year from a tobacco-related disease.¹ Tobacco-related mortality exceeds the number of annual deaths from alcohol, suicide, homicide, motor vehicle crashes, illegal drugs, and HIV infection combined.

In New York City, over 1 million adult residents are current smokers.² In 2005, smoking accounted for approximately 10 000 deaths of New Yorkers; smoking costs the city $4.7 billion annually in direct and indirect costs.²

Smoking prevalence exhibits a strong socioeconomic gradient.³ In 2007, 26.7% of US adults without a high school diploma smoked, compared with 6.6% of individuals with a graduate degree.⁴ Similarly, 30.6% of adults living below the poverty level smoked, compared with 20.4% at or above the poverty level.⁴ The reasons for this disparity are likely multifactorial and may reflect decreased health literacy, reduced access to medical care, lower demand for nicotine addiction treatment, and disparities in the use of cessation aids.^5,6

Effective treatments for tobacco dependence include counseling and pharmacotherapy.⁷ Drug therapy includes nicotine replacement therapy (NRT) products, such as patches, gum, lozenges, inhalers, and prescription tablets such as bupropion and varenicline. NRT is available via prescription or over-the-counter in the form of patches, gum, and lozenges. Over-the-counter pharmacotherapy, which has been available for 2 decades, is safe and effective.⁸ Although New York State Medicaid pays for over-the-counter NRT, physicians and patients are often unaware that this coverage exists.⁹ Thus, access to nonprescription NRT remains an important, if underused, source of treatment of smokers. This underuse of effective, evidence-based treatment may account for some of the disparity in smoking prevalence across income and education.

To evaluate these issues, we studied the availability of nonprescription NRT products in New York City pharmacies and used logistic regression to analyze this availability by the racial, ethnic, and socioeconomic profiles of the communities surrounding the pharmacies. We also studied the availability of cigarettes in these pharmacies, although noting that this would not provide a complete model of overall neighborhood availability of the product.

METHODS

In January through July 2006, we surveyed 30% of all New York City pharmacies, selected at random from the State Department of Health Registry of Retail Pharmacies. Research assistants were instructed to record information regarding pharmacy type (e.g., large chain versus independent) and the availability of tobacco products and nonprescription NRT, product placement, and the presence of advertisements within the store for either. To check interrater reliability, 10% of all pharmacies were visited independently by 2 surveyors.

Pharmacy Site Selection

Registered retail pharmacies located in New York City were identified by use of a master list obtained from the New York State Department of Education, Office of the Professions. The list contained 2127 establishments, which were sorted by New York City borough (Bronx, Brooklyn, Queens, Manhattan, and Staten Island). To generate an allocation schedule of pharmacies, a random sample of pharmacies in New York City were selected from each borough in proportion to the number of its pharmacies by use of an online random plan generator (available at: http://www.randomization.com).

Pharmacies were categorized a priori as either: (1) traditional chain (pharmacies clearly identified with others of the same name such as Rite-Aid, Duane Reade, CVS, Genovese, and Walgreens), (2) independently owned (pharmacies not identified with any other pharmacy or clinic), (3) clinically affiliated (located on the grounds of a medical building, office, or clinic), (4) grocery store, or (5) mass merchant (e.g., Walmart, Target).

Site Surveys

All coinvestigators attended a 3-hour training session that included field visits to 2 pharmacies. Coinvestigators' survey results were then compared against S. L. B's responses, which were used as the criterion standard. All coinvestigators, after having successfully completed the training, then trained 2 to 4 surveyors in their own boroughs by use of the same method. After completion of training, the surveyors were permitted to begin visits to study pharmacies.

Fidelity to protocol was assessed by having 2 surveyors independently sample 10% of all pharmacies in each borough, with measures of agreement calculated for each of the 2 primary survey questions: “Does pharmacy sell NRT?” and “Does pharmacy sell cigarettes?” Proportionate agreement and a Cohen κ calculated on these measures yielded excellent measures of agreement (0.95 ; κ = 0.77 and 0.97; κ = 0.97, respectively).¹⁰

Surveyors conducted 5- to 10-minute walkthroughs of the pharmacy to record, on case report forms, data including pharmacy type (e.g., large chain versus independent), sale of nonprescription nicotine replacement medications, location of NRT, sale of cigarettes, visibility of cigarettes, location of cigarettes, sale of other tobacco products, cigarette advertising, and location of advertisements. (The survey instrument used was adapted from that of Eule et al.¹¹ and is available as a supplement to the online version of this article at http://www.ajph.org.)

Neighborhood Characteristics

Geographic information systems analysis was used to determine the socioeconomic status (SES) of the neighborhood surrounding the pharmacy. Census block data were obtained from the US Census Bureau.¹² Street addresses for each pharmacy were geocoded and entered into ArcGIS 9.0 geographic information systems software (Environmental Systems Research Institute Inc, Redlands, CA). These data were then joined with the spatial boundary file of the census block groups in New York City. A centroid shape file was created for each block group, reducing each to a point, which was approximately the geometric center of the original polygon. Therefore, each block group centroid received the sociodemographic attributes of its associated block group.

The pharmacy data were geocoded to the corresponding borough street file (LotInfo locator file, Space Track, Inc, as derived from New York City Department of Finance, Real Property Attribute Data) by the address field. Quarter-mile radius buffers (to approximate convenient walking distance) were then created for each pharmacy location. The attributes of the pharmacy at the center of its buffer were assigned to the block group centroids captured within that buffer, resulting in a combined database that has each census block group with sociodemographic data and pharmacy data. The block group data were then summarized by unique pharmacy ID numbers, which resulted in a single record for each pharmacy.

Data Analysis

Variables were reported by using means, medians, and ranges. Simple 2-group comparisons were performed with the χ², t, or Mann–Whitney–Wilcoxon tests as appropriate. Multivariate analysis was performed with logistic regression modeling by using a forward conditional approach for categorical data andstepwise linear regression for continuous data. Bivariate binary logistic regressions were performed to compare the sales of cigarettes and the sale of NRT (dependent variables) with the demographic information (predictor variables). Variables were chosen for inclusion in the models based on an F value of less than 0.05. Collinearity was assessed by tolerance level (0.1); collinear variables were removed from the multivariate models. All analyses were performed with SPSS 13.0 (SPSS Inc, Chicago, IL). The α level was set at .05; all tests were 2 sided.

RESULTS

We surveyed a total of 646 pharmacies. The location of the pharmacies studied and their sales of cigarettes and NRT are shown in Figure 1. Summary data from all pharmacies surveyed are shown in Table 1. Of all pharmacies, 89.6% sold any form of NRT and 46.3% sold cigarettes. The race, ethnicity, and SES of the New York City neighborhoods surveyed, using 2000 Census data,¹² are shown in Figure 2.

FIGURE 1.

Availability of cigarettes and nicotine replacement therapy (NRT) in selected New York City pharmacies: January to June, 2006.

Note. Cigs+ = cigarettes are available; Cigs− = no cigarettes are available; NRT+ = NRT is available; NRT− = no NRT available.

TABLE 1.

Characteristics of Pharmacies (N = 646) Surveyed: New York, NY, January–July, 2006

Characteristic	% or Median (IQR)
Type of pharmacy
Chain	34.2
Independent	59.8
Grocery store/market	2.6
Wholesale store	0.6
Clinically affiliated store	2.8
Tobacco products sold
Yes	46.9
No	53.1
Nonprescription NRT sold
Yes	90.8
No	9.2
Tobacco products advertised
Yes	42.8
No	57.2
Location of tobacco product advertisementa
Visible from outside of store	10.4
Inside store, part of cigarette display	94.1
Inside store, separate from cigarette display	7.4
Type of NRT solda
Patch	88.9
Gum	76.0
Lozenge	50.2
NRT advertised
Yes	13.2
No	86.8
Location of NRT advertisementa
Visible from outside of store	4.0
Inside store, part of cigarette display	22.7
Inside store, separate from cigarette display	84.0
Price of cheapest NRT	$31.89 ($23.99–$40.99)
Location of cigarettesa
Behind the front cash register counter	91.9
Behind a separate counter	6.4
Other	2.4
Location of NRTa
Behind the pharmacy counter	72.4
Behind the cashier's counter	35.3
Over-the-counter	8.4
On a shelf, but in a locked case	5.2
Immediately adjacent to cigarettes	10.0

Note. NRT = nicotine replacement therapy; IQR = interquartile range.

^aBecause of advertising or product placement, percentages may add to more than 100%.

FIGURE 2.

Race, ethnicity, and socioeconomic status of New York City neighborhoods, by (a) non-Hispanic Whites, (b) non-Hispanic Blacks, (c) Hispanics, (d) high school graduation rates, and (e) percentage below the poverty level.: 2000.

Source. Bureau of the Census.¹²

The median price of the cheapest nonprescription NRT patch product was $31.89 (interquartile range = $23.99–$40.99). Most patches came in boxes of 14 (54.7%); the remainder came in boxes of 7.

Our main objective was to determine whether the sale of NRT products in pharmacies varied by neighborhood racial composition and SES. Each predictor variable was tested individually to eliminate potential multicollinearity as well as to see the individual relationships more clearly. The results of the logistic models are provided in Table 2.

TABLE 2.

Logistic Regression of NRT and Cigarette Availability in Selected Pharmacies, by Neighborhood SES and Racial Composition: New York, NY, January–July, 2006

	OR (95% CI)	P
NRT sold
% above federal poverty level	1.039 (1.018, 1.059)	<.01
% with high school diploma	1.033 (1.016, 1.051)	<.01
% non-Hispanic White	1.011 (1.001, 1.020)	.02
% non-Hispanic Black	0.991 (0.982, 1.000)	.04
% Hispanic	0.993 (0.983, 1.004)	.20
Cigarettes sold
% above federal poverty level	1.060 (1.044, 1.075)	<.01
% with high school diploma	1.046 (1.034, 1.058)	<.01
% non-Hispanic White	1.018 (1.013, 1.024)	<.01
% non-Hispanic Black	0.987 (0.980, 0.993)	<.01
% Hispanic	0.980 (0.973, 0.987)	<.01

Note. NRT = nicotine replacement therapy; SES = socioeconomic status; OR = odds ratio; CI = confidence interval.

The results for NRT supported the hypothesis that race, education, and income of neighborhoods play a potentially important role in the accessibility or availability of NRT. For example, there was an additional 3.9% likelihood of NRT being available in more-affluent neighborhoods and a 3.3% increased likelihood in neighborhoods in which the residents had a higher education level. Conversely, less-affluent areas or areas with lower educational levels exhibited inverted associations.

The relationships between the demographics and tobacco and NRT were positively correlated to the percentage of individuals who were not below the poverty level as well as the percentage of individuals with a high school diploma living in the neighborhood. The associations between tobacco and NRT sales and the race/ethnicity of neighborhoods were weaker but consistent between the 2 dependent variables. Block groups with a higher percentage of non-Hispanic White residents were positively correlated with greater availability of NRT and cigarettes, whereas areas with higher percentages of non-Hispanic Black and Hispanic residents were negatively correlated. Additionally, all of the regression variables, with the exception of the model for the correlation of the percentage of Hispanic residents with NRT, showed significance (P < .05).

NRT was more expensive in poorer neighborhoods. Pharmacies were dichotomized at the median poverty level (20.18%) of the surrounding neighborhoods. Pharmacies in poorer neighborhoods sold the cheapest NRT at a median price of $34.43 per package, compared with $29.99 at pharmacies in wealthier neighborhoods (P = .04, by the Mann–Whitney–Wilcoxon test). Race, ethnicity, and education level were not associated with price disparities.

Both NRT and cigarettes were more likely to be advertised in pharmacies located in wealthier neighborhoods. Dichotomized at the median level of poverty, pharmacies in poorer neighborhoods advertised NRT and cigarettes in 10.1% and 28.5% of all stores, respectively, compared with 16.3% and 57.7%, respectively, in pharmacies in wealthier areas (P = .03 and P < 0.001, respectively, by the χ² test). Race, ethnicity, and education level were not associated with disparities in advertising these products.

The differences in pricing and advertising were explained by neighborhood variation in the type of pharmacy present. Chain pharmacies were less likely to be located in poorer, less-educated neighborhoods with fewer White residents. The proportion of residents living below the federal poverty level, lacking a high school diploma, or being Black or Hispanic in the neighborhoods in which nonchain pharmacies were located compared with ones in which chain pharmacies were located, respectively, was 24.9% versus 17.5%, 33.0% versus 23.2%, and 51.7% versus 37.0%, respectively (all P < .001).

DISCUSSION

We found a positive association between the availability of NRT and neighborhoods predominantly with White residents and a similar relationship for cigarette availability. The relationships among a pharmacy's availability of NRT or tobacco products and racial, ethnic, educational, and economic variables of the neighborhoods in which the stores were located were not strong, however. We also found that the NRT available in poorer neighborhoods was more expensive than that available in wealthier neighborhoods.

The greater availability of NRT in wealthier, better-educated, predominantly White neighborhoods largely reflected the greater penetration of chain pharmacies in these areas. For example, 216 of 221 (97.7%) of the pharmacies in these areas sold NRT, versus 339 of 384 (88.2%) of the independent pharmacies that were more prevalent in lower-SES areas (P < .001, by the χ² test).

Cigarettes are widely available at a variety of retail outlets, including pharmacies, liquor stores, convenience stores, gas stations, supermarkets, mass merchandisers, and tobacco specialty stores.^13,14 Nicotine replacement products, however, can only be purchased in registered pharmacies. In 1971, the American Pharmaceutical Association recommended that tobacco products not be sold in pharmacies.¹⁵ Since this guideline was issued, however, multiple studies have documented high rates of cigarette sales and merchandising in retail pharmacies, including 100% in large chain pharmacies.^{11, 16-18} A 2004 survey of San Francisco, California, pharmacies found that 78% sold over-the-counter nicotine replacement medications, whereas only 44% of independently owned pharmacies did. As noted by 3 former surgeons general, in the United States, it is often more difficult to purchase nicotine replacement products than tobacco.¹⁹

In 2005, the cigarette industry spent almost $13.11 billion, more than $36 million per day, on advertising and promotion.²⁰ Barbeau et al.²¹ used internal tobacco industry documents to report how 2 of the largest tobacco companies specifically target low-income, less-educated, working class individuals. Other studies have noted that storefront tobacco advertising is far more prevalent in predominantly minority, low-income communities than in nonminority, higher-income communities.²² Additional relevant disparities between White and non-White neighborhoods have been noted in the stocking of narcotic pain medication in pharmacies.²³ Morrison et al.²³ found that pharmacies in non-White neighborhoods were less likely to stock adequate amounts of narcotic analgesics than were pharmacies in White neighborhoods, whereas only 50% of all surveyed pharmacies had enough pain medication in stock to treat someone in severe pain. Although such socioeconomic disparities are prevalent, no study until ours had examined the differences in sales of nicotine replacement medication in communities, stratified by ethnic, racial, and SES.

Most pharmacists do not believe that they should sell cigarettes.¹⁷ Nonetheless, the sale of cigarettes in pharmacies remains high, particularly among the chains, largely because of economic considerations.^17,24 Tobacco sales increase pharmacy revenue, and tobacco companies may provide additional economic incentives based on sales volume and advertisements.²⁴ It therefore appears likely that, for the foreseeable future, at least some pharmacies will continue to sell tobacco products. Of note, the city of San Francisco recently passed legislation banning the sale of tobacco products in pharmacies, but not in grocery stores or “big box” stores that also sell pharmaceuticals.²⁵

It is possible that the pricing structure of nonprescription NRT inhibits potential buyers, although we did not conduct interviews with smokers to assess this. This may account for some of the reported lower demand for tobacco dependence treatment, and lower successful quit rates, in smokers of lower SES or from racial/ethnic minorities.⁶ These issues await further study.

Just as single cigarette packs approximate a 1-day supply, there is interest in making NRT available in similar fashion. In an attempt to expand access to NRT, in January 2008, the health commissioner of the State of New York circulated a petition to the Food and Drug Administration urging wider availability of NRT. The petition called for the Food and Drug Administration (1) to allow over-the-counter NRT to be sold in all retail locations where cigarettes are sold, including convenience stores, gas stations, tobacco specialty stores, grocery stores, and other retail businesses that sell tobacco; (2) to allow over-the-counter NRT to be sold in “daily” units (containing an amount of NRT that would typically be consumed in a 24-hour period) at prices competitive with one 20-count pack of cigarettes and to allow over-the-counter NRT to be advantageously positioned near cigarettes and other tobacco products; and (3) to modify labeling requirements to fully disclose to smokers the benefits of over-the-counter NRT use relative to continued cigarette use, including explanation of the risks associated with each.

Because cigarettes are available from many retail outlets other than pharmacies, our study does not provide a useful model of cigarette or other tobacco product availability. To further evaluate these disparities, it would be necessary to sample the availability of tobacco products in other types of retail stores, such as small markets, gas stations, and delis. This may be a subject for future work.

To our knowledge, this is the first study that attempted to survey the availability of tobacco products and NRT in pharmacies stratified by the racial, ethnic, and SES characteristics of the surrounding neighborhoods. We found that weak but statistically significant racial and socioeconomic disparities existed in access to nonprescription NRT in New York City pharmacies. NRT costs were about $4 more per package in poorer neighborhoods. Cigarettes were also more available in pharmacies in neighborhoods with a higher SES or a higher proportion of Whites, but the model did not account for access to cigarettes from all types of retail outlets. These results may explain some of the excess prevalence of cigarette use in areas of low SES.