Abstract
Introduction:
The use of mail-order pharmacies is generally associated with lower healthcare costs and improved medication adherence. In order to promote the use of mail-order pharmacies, it is important to understand time trends in their use and whether these trends vary by population subgroups.
Methods:
This study used the 1996–2018 Medical Expenditure Panel Survey to determine the annual prevalence of mail-order pharmacy use (defined as purchasing ≥1 prescription from a mail-order or online pharmacy) among U.S. adult prescription users and its variation by population characteristics. Logistic regression was used to determine the correlates of mail-order pharmacy use. Results were presented for medications and therapeutic classes most commonly purchased by mail-order pharmacy exclusive users. Analyses were conducted in December 2020.
Results:
The annual prevalence of mail-order pharmacy use among U.S. adult prescription users increased from 10.2% (95% CI=9.3%, 11.1%) in 1996 to 17.0% (95% CI=15.9%, 18.1%) in 2005, then declined to 15.7% (95% CI=14.9%, 16.6%) by 2018. Absolute differences in the prevalence of use by race/ethnicity, education, and health insurance coverage widened over time, whereas they remained stable when stratifying by sex, age, marital status, region, limitations in daily activities or routine functions, pain interference, health status, number of chronic conditions, and access to medical care. Among mail-order pharmacy exclusive users, the 3 most commonly purchased medications were atorvastatin (16.7%), levothyroxine (13.6%), and lisinopril (13.1%); the 3 most commonly purchased therapeutic classes were cardiovascular agents (57.9%), metabolic agents (52.1%), and central nervous system agents (29.6%).
Conclusions:
The prevalence of mail-order pharmacy use has declined in recent years and displayed significant variation across some population subgroups. Future research should examine whether the declining trends and variation in use may influence the management of chronic conditions and disparities in health and healthcare costs.
INTRODUCTION
Mail-order pharmacies are commonly used for dispensing outpatient prescriptions. Their potential market is enormous owing to the focus on delivering maintenance prescriptions1—the majority of outpatient prescription sales in the U.S.2 Mail-order pharmacies provide patients with numerous benefits, including the convenience of home delivery, online ordering for refills, and a longer days’ supply than at community pharmacies. Compared with community pharmacies, the total costs at mail-order pharmacies are often lower for patients but are more expensive for health plans.3 Health plans’ greater costs are in part due to lower copayments at mail-order pharmacies and patients’ higher rates of medication adherence and refills.3,4 However, medication adherence is typically associated with lowering overall healthcare costs and improving health outcomes for mail-order users,4,5 which justifies higher dispensing costs for health plans. These perceived benefits have, in part, led to a substantial growth in the market share of mail-order pharmacies in the outpatient prescription market, from 6% in 1989 to 15.2% in 2013.6,7
Yet, mail-order pharmacy use has slowed down and even declined in recent years, accounting for 9.5% of the outpatient prescriptions dispensed in 2019.2 Part of this pattern reflects, among others, the growing competition from community pharmacies, the declining advantages of mail-order pharmacies as more employers allow community pharmacies to fill 90-day prescriptions for maintenance medications, policies that favor community pharmacies, and patients’ concerns about the loss or delay of their mail orders.8–11 Furthermore, mail-order pharmacy use varies significantly by SES and healthcare access.12 Non-Hispanic Whites, high-income patients, older adults, college graduates, and individuals with health insurance are more likely to use mail-order than their counterparts.12 These disparities in use may reflect discrepancies in health, healthcare access, and health literacy.13–15 Given the previously documented benefits of mail-order pharmacies for medication adherence,4,5,16 the declining trends and variation in use may influence the management of chronic conditions and disparities in health and healthcare costs.
In order to promote the use of mail-order pharmacies, policymakers and clinicians need a comprehensive understanding of how their use has changed over time and whether these time trends vary across population characteristics. Several studies have estimated the prevalence of use, but they are either outdated,12,17 unrepresentative of the U.S. population,17–19 or unable to assess variation in time trends by detailed population characteristics.2,18 This study uses a nationally representative survey to determine: (1) the prevalence of mail-order pharmacy use from 1996 to 2018 among U.S. adult prescription users, (2) how time trends vary by population subgroups, and (3) which medications are most commonly purchased by mail-order pharmacy exclusive users.
METHODS
Study Sample
The Medical Expenditure Panel Survey Household Component (MEPS) is a nationally representative survey of the civilian non-institutionalized U.S. population drawn from a subsample of households that responded to the prior year’s National Health Interview Survey.20 This study used a harmonized version of the MEPS that was made available at the Integrated Public Use Microdata Series at the University of Minnesota.21 MEPS respondents entered the survey annually as members of a unique panel. They were interviewed in person in 5 consecutive rounds spanning over a 2.5-year period. Data were organized as annual files, each of which included members from 2 panels.17 This study used the full-year consolidated files linked to the prescribed medicines and medical conditions files. Analyses were based on adults aged ≥18 years in the pooled 1996–2018 sample who purchased ≥1 prescription in a year (“prescription users” hereafter, n=324,174). After excluding 7,412 respondents with missing data on pharmacy provider for all of their prescriptions, the final sample was 316,762 (weighted n=146,794,255). This study is exempt from IRB approval because MEPS data are in the public domain.
Measures
Information on prescription purchase came from the prescribed medicines files. In each round, respondents supplied the name of any outpatient prescription that they purchased during that round. Each original purchase or refill was treated as a distinct record. Round-specific medication records were aggregated to the calendar year level based on the time of purchase. Respondents were asked to report the type of pharmacy from which each medication was purchased, with the options including mail-order pharmacies; pharmacies located in another store; pharmacies located in an HMO, clinic, or hospital; independent pharmacies; or online pharmacies. Respondents could list multiple pharmacies associated with their prescriptions purchased in a round or over the calendar year. Mail-order pharmacy users were defined as those who purchased ≥1 prescription from a mail-order or an online (“mail-order” hereafter) pharmacy in a year (n=40,131). The remaining respondents (n=276,631) purchased all of their prescriptions in a year from other types of pharmacies.
Information on sociodemographic characteristics, healthcare access, and health came from the full-year consolidated and the medical conditions files. MEPS asked respondents about the following sociodemographic characteristics: age (categorized in this study as 18–44 years, 45–64 years, and ≥65 years), sex, race/ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, non-Hispanic other, and Hispanic), education (less than high school, high school graduate, and college graduate), marital status (married, divorced/widowed/separated, and never married), household poverty status (<200% of the federal poverty threshold, 200%–399%, and ≥400%), region of residence (Northeast, Midwest, South, and West), and health insurance (only private insurance, only Medicare or a combination of only Medicare and private insurance, any other public insurance [Medicaid, Children’s Health Insurance Program, or any government-sponsored programs that pay for medical care], any military insurance [TRICARE, CHAMPUS, or CHAMP-VA], and uninsured every day in a year). Appendix Text 1 provides detailed information on the construction of the health insurance coverage variable. Respondents reported the amount of travel time to their usual place for medical care, with the response options being <15 minutes, 15–30 minutes, 31–60 minutes, and >60 minutes. Respondents rated their general health, with excellent, very good, good, fair, or poor being the response options. Further, respondents indicated whether pain interfered with their normal work last month (none or a bit of the time, and some/most/all of the time). Limitations in daily activities or routine functions (“any limitation” hereafter) were defined as whether a respondent reported any instrumental activities of daily living limitation, activities of daily living limitation, functional limitation, activity limitation, hearing impairment, or vision impairment. The number of chronic conditions (none, 1, 2, and ≥3) was based on a list of self-reported current medical conditions, including hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, asthma, cancer, chronic kidney disease, chronic obstructive pulmonary disease, dementia, diabetes, hepatitis, HIV, and osteoporosis.23 A medical condition was defined as current if it was linked to an event during the interview round (i.e., prescription medication use, outpatient, inpatient, doctor’s office, emergency department, or home health visit) or if a respondent was experiencing the condition during the year.
Statistical Analysis
Descriptive statistics were used to assess the overall prevalence of mail-order pharmacy use among the full sample of 316,762 prescription users. Logistic regression was used to determine the association between mail-order use and respondents’ characteristics. The overall and annual prevalence of use was stratified by sociodemographic characteristics, healthcare access, and health. The samples used to estimate variation in use by these characteristics depended on the degree of missingness of each characteristic. Among mail-order pharmacy exclusive users (who purchased all of their prescriptions in a year from mail-order pharmacies, n=9,979), this study reported the 25 most commonly purchased medications. Because the reported pharmacy type was specific to a respondent, not their individual prescriptions, the authors were not able to calculate the prevalence of mail-order pharmacy use among all prescription users for each individual medication. Similar results were reported for commonly purchased therapeutic classes based on the Multum Lexicon therapeutic classification. A common variance structure designed specifically for pooling samples was used to reflect MEPS’s complex survey design. All analyses were weighted using the final basic annual person weights. Data were analyzed in December 2020 using Stata, version 13.0.
RESULTS
Table 1 presents descriptive statistics for sociodemographic characteristics, healthcare access, and health of the full sample and by mail-order pharmacy use. Overall, 14.7% (95% CI=14.3%, 15.1%) of prescription users purchased ≥1 prescription from a mail-order pharmacy in a year. In Appendix Table 1, compared with the reference groups, the ORs (unadjusted and adjusted) in mail-order pharmacy use were higher (p<0.05) for prescription users who were aged ≥45 years, non-Hispanic White, high school or college educated, married, middle- or high-income, and insured by private insurance, Medicare, or military insurance; who reported less-than-excellent health status, pain interference, any limitation, any chronic conditions; and who had a usual place for medical care (Appendix Table 1 provides the estimated ORs).
Table 1.
Prevalence of Mail-Order Pharmacy Use Among U.S. Adult Prescription Users – 1996–2018
| (1) | (2) | (3) | |
|---|---|---|---|
| Variable | Full sample | No mail-order pharmacy use | Any mail-order pharmacy use |
| Weighted prevalence %a,b (95% CI) |
|||
| Sample size (unweighted), n | 316,762 | 276,631 | 40,131 |
| Population size (weighted), n | 146,794,255 | 125,217,843 | 21,576,411 |
| Used mail-order pharmacies | 14.7 (14.3, 15.1) |
||
| Sociodemographic variables | |||
| Age groups, years | |||
| 18–44 | 39.4 (38.9, 40.0) |
43.4 (42.8, 44.0) |
16.4 (15.7, 17.1) |
| 45–64 | 36.2 (35.9, 36.6) |
35.3 (35.0, 35.7) |
41.4 (40.5, 42.4) |
| ≥65 | 24.3 (23.8, 24.9) |
21.2 (20.8, 21.7) |
42.2 (41.0, 43.3) |
| Sex | |||
| Men | 41.6 (41.4, 41.9) |
41.3 (41.1, 41.6) |
43.6 (43.0, 44.2) |
| Women | 58.4 (58.1, 58.6) |
58.7 (58.4, 58.9) |
56.4 (55.8, 57.0) |
| Race/ethnicity | |||
| Non-Hispanic White | 74.2 (73.3, 75.2) |
72.2 (71.2, 73.3) |
85.8 (85.0, 86.5) |
| Non-Hispanic Black | 10.3 (9.7, 11.0) |
11.1 (10.4, 11.8) |
5.7 (5.2, 6.1) |
| Non-Hispanic Asian | 3.7 (3.3, 4.0) |
3.8 (3.4, 4.2) |
2.9 (2.5, 3.3) |
| Non-Hispanic others | 1.8 (1.6, 2.0) |
1.9 (1.6, 2.1) |
1.3 (1.1, 1.5) |
| Hispanic | 10.0 (9.3, 10.7) |
10.9 (10.2, 11.7) |
4.4 (4.0, 4.8) |
| Educational attainmentc | |||
| Less than high school | 15.0 (14.6, 15.3) |
16.0 (15.6, 16.4) |
9.0 (8.5, 9.4) |
| High school graduate | 31.2 (30.7, 31.7) |
31.2 (30.7, 31.7) |
31.2 (30.3, 32.1) |
| College or above | 53.8 (53.2, 54.5) |
52.8 (52.1, 53.5) |
59.9 (58.8, 60.9) |
| Marital statusd | |||
| Married | 57.0 (56.5, 57.5) |
55.2 (54.6, 55.7) |
67.6 (66.6, 68.5) |
| Divorced, widowed, or separated | 23.1 (22.7, 23.5) |
23.1 (22.7, 23.5) |
23.1 (22.4, 23.9) |
| Never married | 19.9 (19.5, 20.2) |
21.7 (21.4, 22.1) |
9.3 (8.8, 9.8) |
| Household poverty levele | |||
| Low-income | 27.2 (26.6, 27.8) |
28.7 (28.1, 29.3) |
18.2 (17.6, 18.9) |
| Middle-income | 29.6 (29.2, 29.9) |
29.9 (29.5, 30.3) |
27.6 (26.8, 28.4) |
| High-income | 43.2 (42.5, 44.0) |
41.4 (40.6, 42.1) |
54.2 (53.1, 55.3) |
| Region of residencef | |||
| Northeast | 18.6 (17.6, 19.6) |
18.3 (17.2, 19.3) |
20.6 (19.0, 22.2) |
| Midwest | 23.2 (22.1, 24.3) |
23.0 (21.9, 24.1) |
24.0 (22.5, 25.5) |
| South | 36.9 (35.5, 38.3) |
37.2 (35.8, 38.7) |
34.9 (33.2, 36.7) |
| West | 21.3 (20.0, 22.7) |
21.5 (20.1, 22.9) |
20.5 (18.9, 22.1) |
| Health characteristics and access to care | |||
| Self-reported health statusg | |||
| Excellent | 13.6 (13.3, 13.9) |
14.2 (13.9, 14.5) |
10.3 (9.8, 10.9) |
| Very good | 34.9 (34.5, 35.3) |
34.8 (34.4, 35.2) |
35.3 (34.4, 36.1) |
| Good | 33.5 (33.2, 33.8) |
33.0 (32.6, 33.3) |
36.4 (35.7, 37.1) |
| Fair | 14.3 (14.0, 14.6) |
14.2 (13.9, 14.5) |
14.6 (14.0, 15.2) |
| Poor | 3.7 (3.6, 3.8) |
3.8 (3.6, 3.9) |
3.4 (3.2, 3.7) |
| Frequency pain interfered with workh | |||
| None or a bit of the time | 73.4 (73.0, 73.8) |
73.8 (73.4, 74.2) |
71.4 (70.7, 72.2) |
| Some, most, or all of the time | 26.6 (26.2, 27.0) |
26.2 (25.8, 26.6) |
28.6 (27.8, 29.3) |
| Any limitationi | |||
| No | 66.6 (66.1, 67.1) |
68.0 (67.5, 68.5) |
58.8 (57.9, 59.7) |
| Yes | 33.4 (32.9, 33.9) |
32.0 (31.5, 32.5) |
41.2 (40.3, 42.1) |
| Number of chronic conditionsj | |||
| None | 38.9 (38.3, 39.5) |
42.4 (41.8, 43.1) |
17.9 (17.2, 18.6) |
| 1 | 26.4 (26.1, 26.7) |
26.6 (26.3, 26.9) |
25.3 (24.6, 26.0) |
| 2 | 16.0 (15.7, 16.2) |
14.8 (14.6, 15.1) |
22.7 (22.1, 23.3) |
| ≥3 | 18.7 (18.3, 19.2) |
16.1 (15.7, 16.6) |
34.1 (33.1, 35.1) |
| Health insurance | |||
| Private insurance only | 55.1 (54.4, 55.8) |
56.1 (55.4, 56.8) |
49.2 (48.1, 50.4) |
| Only Medicare or a combination of only Medicare and private insurance | 21.6 (21.1, 22.1) |
18.6 (18.2, 19.1) |
38.5 (37.5, 39.6) |
| Any public insurance | 13.1 (12.6, 13.5) |
14.5 (14.0, 15.0) |
4.6 (4.3, 4.9) |
| Any military insurance | 3.2 (3.0, 3.4) |
2.8 (2.6, 3.1) |
5.5 (5.0, 5.9) |
| Uninsured | 7.0 (6.8, 7.3) |
7.9 (7.6, 8.1) |
2.2 (2.0, 2.4) |
| Travel time to usual place of medical carek | |||
| No usual source of medical care | 12.5 (12.1, 12.8) |
13.8 (13.4, 14.2) |
5.6 (5.2, 6.0) |
| <15 minutes | 43.5 (42.8, 44.2) |
43.2 (42.5, 43.9) |
45.0 (43.7, 46.3) |
| 15–30 minutes | 34.7 (34.1, 35.4) |
33.9 (33.3, 34.5) |
39.0 (37.9, 40.2) |
| 31–60 minutes | 7.7 (7.4, 8.0) |
7.6 (7.3, 7.8) |
8.6 (8.1, 9.2) |
| >60 minutes | 1.6 (1.5, 1.7) |
1.5 (1.4, 1.6) |
1.7 (1.4, 1.9) |
Source/Notes: Medical Expenditure Panel Survey 1996–2018.
Weighted estimates using the final basic annual person weights.
Estimated may not add up to 100% due to rounding error.
Excluded 1,775 respondents who had missing data on educational attainment.
Excluded 5 respondents who had missing data on marital status.
Households were classified into 3 categories based on their total income and household composition: low-income (<200% of the federal poverty threshold), middle-income (200%–399%), and high-income (>400%).
Excluded 1 respondent who had missing data on region of residence.
Excluded 65,199 respondents who had missing data on health status. MEPS started collecting this variable in 2000 as part of the Self-Administered Questionnaire (SAQ) portion of the survey. Respondents interviewed before 2000 were removed when analyzing this variable.
Excluded 64,954 respondents who had missing data on pain interference. MEPS started collecting this variable in 2000 as part of the Self-Administered Questionnaire (SAQ) portion of the survey. Respondents interviewed before 2000 were removed when analyzing this variable.
Excluded 2,583 respondents who had missing data on limitation status. The variable was constructed based on whether or not a respondent reported any instrumental activities of daily living (IADL) limitations, activities of daily living (ADL) limitations, functional limitations, activity limitations, hearing impairment, or vision impairment.
Excluded 43,896 respondents who had missing data on chronic conditions. In 2016, MEPS started using the ICD-10 instead of the ICD-9. As such, the medical classifications of conditions in 2016 onward are still under review. Respondents interviewed in 2016 onward were removed when analyzing this variable. The following conditions were considered chronic in this study: hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, asthma, cancer, chronic kidney disease, chronic obstructive pulmonary disease, dementia (Alzheimer’s and other senile dementias), diabetes, hepatitis, HIV, and osteoporosis.
Excluded 87,884 respondents who had missing data on travel time to the usual place of medical care. MEPS started collecting this variable in 2002. Respondents interviewed before 2002 were removed when analyzing this variable.
Appendix Figure 1 demonstrates trends in mail-order pharmacy use among prescription users from 1996 to 2018. The prevalence of use increased from 10.2% (95% CI=9.3%, 11.1%) in 1996 to 17.0% (95% CI=15.9%, 18.1%) in 2005, stagnated between 2006 and 2011, then declined to 15.7% (95% CI=14.9%, 16.6%) in 2018. Age-standardized estimates (calculated based on the standard population from the 2000 Census data) were slightly lower than crude estimates, although the patterns over time remained unchanged.
Figure 1 depicts time trends in mail-order pharmacy use among prescription users by sex, age, race/ethnicity, and income. The sex difference in use was relatively small over time. The prevalence of use was low among adults aged 18–44 years throughout the study period, whereas it increased rapidly between 1996 and 2006 among adults aged ≥45 years, declining thereafter. By race/ethnicity, the disparity in use widened over time. Between 1996 and 2018, absolute differences in the prevalence of use between non-Hispanic White compared with non-Hispanic Black, non-Hispanic Asian, and Hispanic adults grew 3.6 percentage points (95% CI=1.0%, 6.2%), 6.4 percentage points (95% CI=1.2%, 11.5%), and 5.0 percentage points (95% CI=2.3%, 7.7%), respectively. When stratifying by household income, disparity in use generally declined since 2008. Between 2008 and 2018, discrepancies in the prevalence of use between high-income compared with middle- and low-income respondents diminished by 4.9 percentage points (95% CI= −8.0%, −1.7%) and 6.8 percentage points (95% CI= −9.6%, −4.0%).
Figure 1.
Trends in mail-order pharmacy use among U.S. adult prescription users by sex, age, race/ethnicity, and household income – 1996–2018.
Source/Notes: Medical Expenditure Panel Survey 1996–2018. All analyses were based on the full sample of 316,762 adult prescription users. Household income (Panel D) was categorized into 3 mutually exclusive groups based on their total income and household composition: low-income (<200% of the federal poverty threshold), middle-income (200%–399%), and high-income (>400%).
NH, non-Hispanic.
The prevalence of mail-order pharmacy use among other subpopulations generally increased during the first 10 years of the study period, stagnated from 2006 to 2010, and declined thereafter. Disparities in use grew for some population characteristics. The gap in mail-order use between prescription users with a college degree and those without a high school diploma increased 7.8 percentage points (95% CI=5.2%, 10.4%) between 1996 and 2018 (Figure 2). Over the same period, disparities in use grew 28.2 percentage points (95% CI=20.0%, 35.9%) and 3.8 percentage points (95% CI=1.1%, 6.5%) for prescription users with military insurance and private insurance, respectively, compared with uninsured respondents (Appendix Figure 2). The gap in use grew modestly by 5.9 percentage points (95% CI=1.9%,9.9%) from 1996 to 2018 between prescriptions users with and without hearing difficulties (Appendix Figure 3). By contrast, absolute differences in the prevalence of use have been relatively stable over time when stratifying by marital status, region, any limitation, pain interference, health status, number of chronic conditions, and the amount of travel time to a usual place for medical care (Figure 2 and Appendix Figures 2 and 3).
Figure 2.
Trends in mail-order pharmacy use among U.S. adult prescription medication users by marital status, education, and region – 1996–2018.
Source/Notes: Medical Expenditure Panel Survey 1996–2018. Panel A excluded 5 respondents from the full analytic sample of 316,762 prescription users due to missing data on marital status. Panel B excluded 1,775 prescription users with missing data on educational attainment. Panel C excluded 1 prescription user with missing information on region of residence.
Among mail-order pharmacy exclusive users, most commonly purchased medications were intended for treatment or prevention of high cholesterol levels, hypothyroidism, high blood pressure, cardiovascular disease, diabetes, infections, hypokalemia, osteoporosis, blood clots, pain, and mental health disorders. Some of these medications included atorvastatin (16.7%, 95% CI=15.6%, 17.8%), levothyroxine (13.6%, 95% CI=12.5%, 14.7%), lisinopril (13.1%, 95% CI=12.1%, 14.1%), simvastatin (13.1%, 95% CI=12.1%, 14.0%), and metoprolol (10.7%, 95% CI=9.8%, 11.7%) (Table 2 and Appendix Table 2). In addition, most commonly purchased therapeutic classes among mail-order pharmacy exclusive users included cardiovascular agents (57.9%, 95% CI=56.5%, 59.3%), metabolic agents (52.1%, 95% CI=50.6%, 53.7%), central nervous system agents (29.6%, 95% CI=28.3%, 31.0%), and hormones or hormone modifiers (29.6%, 95% CI=28.2%, 31.0%) (Appendix Table 3).
Table 2.
Top 25 Prescription Medications Purchased by Mail-Order Pharmacy Exclusive Users – 1996–2018
| Medication | Among all mail-order pharmacy exclusive users, weighted percent of users who purchased the prescription medication |
|---|---|
| (95% CI)a | |
| N (unweighted)=9,932a | |
| Atorvastatin | 16.7 (15.6, 17.8) |
| Levothyroxine | 13.6 (12.5, 14.7) |
| Lisinopril | 13.1 (12.1, 14.1) |
| Simvastatin | 13.1 (12.1, 14.0) |
| Metoprolol | 10.7 (9.8, 11.7) |
| Metformin | 9.4 (8.5, 10.3) |
| Amlodipine | 8.6 (7.8, 9.4) |
| Atenolol | 7.6 (6.7, 8.4) |
| Hydrochlorothiazide | 7.2 (6.5, 7.9) |
| Omeprazole | 7.2 (6.3, 8.0) |
| Furosemide | 4.5 (3.9, 5.1) |
| Pravastatin | 3.9 (3.3, 4.5) |
| Losartan | 3.9 (3.3, 4.4) |
| Azithromycin | 3.8 (3.4, 4.3) |
| Potassium chloride | 3.8 (3.2, 4.3) |
| Alendronate | 3.7 (3.1, 4.3) |
| Rosuvastatin | 3.6 (3.0, 4.3) |
| Clopidogrel | 3.6 (3.1, 4.1) |
| Conjugated estrogens | 3.4 (2.9, 3.9) |
| Warfarin | 3.4 (2.9, 3.9) |
| Acetaminophen-hydrocodone | 3.4 (2.9, 3.9) |
| Sertraline | 3.2 (2.7, 3.7) |
| Albuterol | 3.2 (2.7, 3.6) |
| Tamsulosin | 3.1 (2.5, 3.6) |
| Glipizide | 3.0 (2.5, 3.5) |
Source/Notes: Medical Expenditure Panel Survey 1996–2018.
The denominator is the number of U.S. adult prescription users who exclusively used mail-order pharmacies to purchase all of their prescriptions (mail-order pharmacy exclusive users) in a year. The numerator is the number of mail-order pharmacy exclusive users who purchased the prescription medication.
Out of 40,131 respondents who reported any use of mail-order pharmacies in a calendar year, 9,979 exclusively used mail-order pharmacies to purchase all of their prescriptions. Of these mail-order pharmacy exclusive users, 47 respondents who had missing information on the names of all of their reported prescriptions were removed. The final sample was 9,932 mail-order pharmacy exclusive users.
DISCUSSION
To the authors’ knowledge, this study is the first to use a nationally representative survey of U.S. prescription users to assess the prevalence of mail-order pharmacy use over the past 2 decades. The prevalence of use among prescription users increased from 10.2% in 1996 to 17.0% in 2005, then declined to 15.7% by 2018. A large portion of mail-order pharmacy exclusive users purchased cardiovascular agents, metabolic agents, central nervous system agents, hormones or hormone modifiers, and gastrointestinal agents, suggesting that the use of mail-order pharmacies was often intended for treatment or prevention of chronic conditions.1 Although antibiotics are frequently used for treatment of acute infection and are typically purchased at community pharmacies, they have been increasingly purchased via mail order for treatment of chronic acne and other indicators.24,25
The prevalence of mail-order pharmacy use has stagnated and then declined in recent years, despite the rapid increase before 2006. Medicare Part D might partially be responsible for this pattern.10 Since 2006, Part D plans that offered prescription benefits at mail-order pharmacies must provide the same benefits at retail pharmacies. Given that 83.7% of older adults opposed mail-order pharmacies (if they could cause local pharmacies to close), this requirement likely was an important reason why many Medicare beneficiaries opted against mail-order prescription benefits.8 Moreover, many chain pharmacies introduced low-cost generic programs in 2006 that offered a month’s supply of many medications for $4.26 Such efforts potentially diminished the economic advantages of mail-order pharmacies. Furthermore, fewer employers required mail-order services for filling 90-day prescriptions for maintenance medications, with a decline from 45% in 2011 to 23% in 2012.11 This shift likely explained a decline in mail-order pharmacy use after 2011, especially among privately insured adults.
The prevalence of mail-order pharmacy use varied by population subgroups. Non-Hispanic White, high- and middle-income, college educated, and married prescription users were more likely to use mail-order than their peers, potentially owing to, among other factors, their higher levels of health literacy and the use of technologies to access health information.14,27,28 Older, Medicare-insured, disabled, and chronically ill adults were more likely to purchase prescriptions via mail than their counterparts, possibly because they consumed more maintenance medications and might have physical limitations that prevented them from visiting community pharmacies.13,29 Furthermore, the prevalence of use was higher among adults covered by private and military insurance compared with other insurance types, perhaps owing to these systems’ requirements for using mail services to fill 90-day prescriptions and a potentially lower copayment at mail-order pharmacies.3,7,11,16 Finally, having a usual place for medical care (regardless of travel time to the facility) was associated with an increase in mail-order pharmacy use, in part because most patients needed a prescription to place a mail order.
The growing disparities in mail-order pharmacy use by race/ethnicity and education are potentially due to various factors. First, there are persistent gaps in the overall health insurance coverage by these characteristics across time.15,30 Second, the racial/ethnic and education gaps in the sources of health insurance coverage for mail-order prescriptions have remained relatively unchanged over time. The present study showed that privately insured adults were more likely to use mail order than those with public insurance or without any insurance. Yet, the absolute differences in the prevalence of having private insurance coverage between non-Hispanic White and non-Hispanic Black, non-Hispanic White and Hispanic, and college graduates and adults without a high school diploma remained around 21, 28 and 52 percentage points between 2008 and 2014.15,30 Finally, the growing or persistent disparities in access to physician care, SES, and the use of technologies to access health information, among other factors, may depress the use of mail-order pharmacies for racial/ethnic minorities and adults with low educational attainment.31,32 Future studies should investigate the implications of growing racial/ethnic and education disparities in mail-order pharmacy use for the widening discrepancies in medication adherence and health outcomes.33,34 They should also examine ways to reduce disparities in such use.
The declining trend in mail-order pharmacy use is likely to be reversed during the coronavirus disease 2019 (COVID-19) pandemic. In the last week of March 2020, prescriptions dispensed via mail increased by 21% relative to the previous year, suggesting that many patients had switched to mail-order pharmacies to avoid potential exposure in stores.35 Findings from this study provided policymakers and clinicians with baseline estimates to predict the use of mail-order pharmacies during the COVID-19 pandemic. Many older adults, socioeconomically advantaged individuals, and chronically ill patients will likely continue to rely on mail-order pharmacies during the COVID-19 pandemic to avoid the potential risk for infection in stores. The pandemic may also change the preferred dispensing method for many population subgroups that have a historically low prevalence of mail-order use such as racial/ethnic minorities, young adults, never-married individuals, low-income adults, and those without a high school diploma.
Moreover, recent operational changes at the U.S. Postal Services and the resulting slowdown in mail delivery have raised concerns among those who rely on mail order for their lifesaving prescriptions.36,37 Given that many mail-order pharmacy users are socioeconomically advantaged, it is likely that they will manage to switch to an alternative dispensing outlet if any postal delay occurs. However, some older adults and chronically ill patients may be affected by the slowdown of mail delivery because they are physically least able to travel to a pharmacy to obtain their medications. Future research should investigate the effect of postal service delivery delays for mail-order prescriptions on medication use, medication adherence, and health outcomes.
Limitations
The present study has some limitations. First, MEPS lacks information on over-the-counter medications. Second, the measure of mail-order pharmacy use was based on self-report and thus may be affected by recall bias. However, among prescription users, MEPS was able to contact the pharmacy of more than half of respondents to verify their answers. Third, the varying degrees of missingness in some population characteristics, the declining response rates over time, and changes in the survey design in 2007 and 2017 might influence the observed trends in mail-order pharmacy use.
CONCLUSIONS
The prevalence of mail-order pharmacy use has declined among U.S. prescription users in recent years and displayed significant variation between population subgroups. Future research should examine the consequences of growing disparities in mail-order pharmacy use on medication adherence and health outcomes.
Supplementary Material
ACKNOWLEDGMENTS
Pascal Geldsetzer was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number KL2TR003143. The funder had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
No financial disclosures were reported by the authors of this paper.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Pharmacy Benefit Managers: Ownership of Mail-Order Pharmacies. Washington, DC: Federal Trade Commission; 2005. https://www.ftc.gov/reports/pharmacy-benefit-managers-ownership-mail-order-pharmacies-federal-trade-commission-report. Accessed March 11, 2021. [Google Scholar]
- 2.Medicine Spending and Affordability in the United States. IQVIA Institute for Human Data Science; 2020. https://www.iqvia.com/insights/the-iqvia-institute/reports/medicine-spending-and-affordability-in-the-us?fbclid=IwAR1yC55Ib1CHHXimCe7rJdxwDD1q2yKt9fdwE1IeXJ8gqDp5rqbnTuMHaMo. Accessed August 22, 2020. [Google Scholar]
- 3.Carroll NV, Brusilovsky I, York B, Oscar R. Comparison of costs of community and mail service pharmacy. J Am Pharm Assoc (2003). 2005;45(3):336–343. 10.1331/1544345054003796. [DOI] [PubMed] [Google Scholar]
- 4.Devine S, Vlahiotis A, Sundar H. A comparison of diabetes medication adherence and healthcare costs in patients using mail order pharmacy and retail pharmacy. J Med Econ. 2010;13(2):203–211. 10.3111/13696991003741801. [DOI] [PubMed] [Google Scholar]
- 5.Schmittdiel JA, Karter AJ, Dyer W, et al. The comparative effectiveness of mail order pharmacy use vs. local pharmacy use on LDL-C control in new statin users. J Gen Intern Med. 2011;26(12):1396–1402. 10.1007/s11606-011-1805-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Fein A 2013 Pharmacy Market Analysis: Chains Up, Mail Down. Drug Channels Institute; 2014. https://www.drugchannels.net/2014/05/2013-pharmacy-market-analysis-chains-up.html. Accessed December 7, 2020. [Google Scholar]
- 7.Horgan C, Goody B, Knapp D, Fitterman L. The role of mail service pharmacies. Health Aff (Millwood). 1990;9(3):66–74. 10.1377/hlthaff.9.3.66. [DOI] [PubMed] [Google Scholar]
- 8.Rupp MT. Attitudes of Medicare-eligible Americans toward mail service pharmacy. J Manag Care Spec Pharm. 2013;19(7):564–572. 10.18553/jmcp.2013.19.7.564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fein A. The 2020 Economic Report on U.S. Pharmacies and Pharmacy Benefit Managers. Drugs Channels Institute; 2020. https://drugchannelsinstitute.com/products/industry_report/pharmacy/. Accessed December 6, 2020. [Google Scholar]
- 10.Pub. 100–18 Medicare Prescription Drug Benefit Manual. Department of Health & Human Services. Center for Medicare & Medicaid Services; 2008. https://www.cms.gov/Regulations-and-Guidance/Guidance/Transmittals/downloads/R1PDB.pdf. Accessed August 27, 2020. [Google Scholar]
- 11.Wu J, Davis–Ajami ML, Noxon V. Patterns of use and expenses associated with mail-service pharmacy in adults with diabetes. J Am Pharm Assoc (2003). 2015;55(1):41–51. 10.1331/japha.2015.14058. [DOI] [PubMed] [Google Scholar]
- 12.Ma J, Wang L. Characteristics of mail-order pharmacy users: results from the Medical Expenditures Panel Survey. J Pharm Pract. 2018;33(3):293–298. 10.1177/0897190018800188. [DOI] [PubMed] [Google Scholar]
- 13.Blackwell DL, Lucas JW, Clarke TC. Summary health statistics for US adults: National Health Interview Survey, 2012. Vital Health Stat 10. 2014;(260):1–161. [PubMed] [Google Scholar]
- 14.Brodie M, Flournoy RE, Altman DE, Blendon RJ, Benson JM, Rosenbaum MD. Health information, the internet, and the digital divide. Health Aff (Millwood). 2000;19(6):255–265. 10.1377/hlthaff.19.6.255. [DOI] [PubMed] [Google Scholar]
- 15.Buchmueller TC, Levinson ZM, Levy HG, Wolfe BL. Effect of the Affordable Care Act on racial and ethnic disparities in health insurance coverage. Am J Public Health. 2016;106(8):1416–1421. 10.2105/ajph.2016.303155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fernandez EV, McDaniel JA, Carroll NV. Examination of the link between medication adherence and use of mail-order pharmacies in chronic disease states. J Manag Care Spec Pharm. 2016;22(11):1247–1259. 10.18553/jmcp.2016.22.11.1247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Phelan J, Ergun D, Langer G, Holyk G. Medication Adherence in America: A National Report Card. National Community Pharmacists Association; 2013. http://www.ncpa.co/adherence/AdherenceReportCard_Full.pdf. Accessed August 22, 2020. [Google Scholar]
- 18.Cubanski J, Biniek J, Rae M, Damico A, Frederiksen B, Salganicoff A. Mail Delays Could Affect Mail-Order Prescriptions for Millions of Medicare Part D and Large Employer Plan Enrollees. https://www.kff.org/coronavirus-covid-19/issue-brief/mail-delays-could-affect-mail-order-prescriptions-for-millions-of-medicare-part-d-and-large-employer-plan-enrollees/. Published August 20, 2020. Accessed August 29, 2020. [Google Scholar]
- 19.Rashrash ME, Tomaszewski DM, Schommer JC, Brown LM. Consumers’ characteristics associated with the use of mail pharmacy services in the United States: findings from the 2015 National Consumer Survey on the Medication Experience. J Am Pharm Assoc (2003). 2017;57(2):206–210. 10.1016/j.japh.2016.09.007. [DOI] [PubMed] [Google Scholar]
- 20.Cohen JW, Cohen SB, Banthin JS. The Medical Expenditure Panel Survey: a national information resource to support healthcare cost research and inform policy and practice. Med Care. 2009;47(7 suppl 1):S44–S50. 10.1097/mlr.0b013e3181a23e3a. [DOI] [PubMed] [Google Scholar]
- 21.Blewett LA, Drew JAR, Griffin R, Williams KC. IPUMS Health Surveys: Medical Expenditure Panel Survey, Version 1.1 [Dataset]; 2019. https://meps.ipums.org. Accessed August 23, 2020. [Google Scholar]
- 22.Agency for Healthcare Research and Quality. MEPS HC-036: 1996–2017 Pooled Linkage Variance Estimation File. Agency for Healthcare Research and Quality; 2019. https://meps.ahrq.gov/data_stats/download_data/pufs/h36/h36u17doc.shtml. Accessed March 11, 2021. [Google Scholar]
- 23.Goodman RA, Posner SF, Huang ES, Parekh AK, Koh HK. Defining and measuring chronic conditions: imperatives for research, policy, program, and practice. Prev Chronic Dis. 2013;10:E66. 10.5888/pcd10.120239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Suda KJ, Hicks LA, Roberts RM, Hunkler RJ, Matusiak LM, Schumock GT. Antibiotic expenditures by medication, class, and healthcare setting in the United States, 2010–2015. Clin Infect Dis. 2018;66(2):185–190. 10.1093/cid/cix773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Barbieri JS, Bhate K, Hartnett KP, Fleming-Dutra KE, Margolis DJ. Trends in oral antibiotic prescription in dermatology, 2008 to 2016. JAMA Dermatol. 2019;155(3):290–297. 10.1001/jamadermatol.2018.4944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Choudhry NK, Shrank WH. Four-dollar generics – increased accessibility, impaired quality assurance. N Engl J Med. 2010;363(20):1885–1887. 10.1056/nejmp1006189. [DOI] [PubMed] [Google Scholar]
- 27.Mantwill S, Monestel-Umaña S, Schulz PJ. The relationship between health literacy and health disparities: a systematic review. PLoS One. 2015;10(12):e0145455. 10.1371/journal.pone.0145455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rikard RV, Thompson MS, McKinney J, Beauchamp A. Examining health literacy disparities in the United States: a third look at the National Assessment of Adult Literacy (NAAL). BMC Public Health. 2016;16:975. 10.1186/s12889-016-3621-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kantor ED, Rehm CD, Haas JS, Chan AT, Giovannucci EL. Trends in prescription drug use among adults in the United States from 1999–2012. JAMA. 2015;314(17):1818–1831. 10.1001/jama.2015.13766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Health Insurance Coverage Type. The State Health Access Data Assistance Center http://statehealthcompare.shadac.org/trend/14/health-insurance-coverage-type-by-education#1/1/5,4,1,10,86,9,8,6,52/1,2,3,4,5,6,7,8,15,24,25/27. Accessed December 12, 2020. [Google Scholar]
- 31.Fareed N, Swoboda CM, Jonnalagadda P, Walker DM, Huerta TR. Differences between races in health information seeking and trust over time: evidence from a cross-sectional, pooled analyses of HINTS data. Am J Health Promot. 2021;35(1):84–92. 10.1177/0890117120934609. [DOI] [PubMed] [Google Scholar]
- 32.Mahmoudi E, Jensen GA. Diverging racial and ethnic disparities in access to physician care: comparing 2000 and 2007. Med Care. 2012;50(4):327–334. 10.1097/mlr.0b013e318245a111. [DOI] [PubMed] [Google Scholar]
- 33.Desai R, Thakkar S, Fong HK, et al. Rising trends in medication non-compliance and associated worsening cardiovascular and cerebrovascular outcomes among hospitalized adults across the United States. Cureus. 2019;11(8):e5389. 10.7759/cureus.5389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bleich SN, Jarlenski MP, Bell CN, LaVeist TA. Health inequalities: trends, progress, and policy. Annu Rev Public Health. 2012;33(1):7–40. 10.1146/annurev-publhealth-031811-124658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hopkins J Mail-Order Drug Delivery Rises During Coronavirus Lockdowns. The Wall Street Journal. May 12, 2020. https://www.wsj.com/articles/mail-order-drug-delivery-rises-during-coronavirus-lockdowns-11589281203. Accessed August 27, 2020. [Google Scholar]
- 36.Axios/Ipsos Poll – Wave 22. Axios; 2020. https://www.ipsos.com/sites/default/files/ct/news/documents/2020-08/topline-axios-w22.pdf. Accessed August 29, 2020. [Google Scholar]
- 37.Rapidly Increasing Postal Service Delivery Delays for Mail-Order Prescription Drugs Pose Health Risks for Millions of Americans. Staff Report Prepared for U.S. Senator Elizabeth Warren and U.S. Senator Bob Casey; 2020. https://www.warren.senate.gov/imo/media/doc/Warren%20Casey%20Mail%20Order%20Drug%20Staff%20Report%20Final.pdf. Accessed December 18, 2020. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.