O. Kenrik Duru, MD, MSHS

Many health programs struggle with low enrollment rates.

To compare the characteristics of populations enrolled in a new 

health plan when employer groups implement voluntary versus 

We analyzed enrollment rates resulting from 2 different strategies: 

voluntary and automatic enrollment. We used regression 

modeling to estimate the associations of patient characteristics 

with the probability of enrolling within each strategy. The subjects 

were 5014 eligible employees from 11 self-insured employers who 

had purchased the Diabetes Health Plan (DHP), which offers free 

or discounted copayments for diabetes related medications, testing 

supplies, and physician visits. Six employers used voluntary 

enrollment while 5 used automatic enrollment. 

The main outcome of interest was enrollment into the DHP. 

Predictors were gender, age, race/ethnicity, dependent status, 

household income, education level, number of comorbidities, and 

Overall, the proportion of eligible members who were enrolled 

within the automatic enrollment strategy was 91%, compared 

with 35% for voluntary enrollment. Income was a significant predictor 

for voluntary enrollment but not for automatic enrollment. 

Within automatic enrollment, covered dependents, Hispanics, 

and persons with 1 nondiabetes comorbidity were more likely to 

Employer group was also a significant correlate of enrollment. 

Notably, all demographic groups had higher DHP enrollment 

rates under automatic enrollment than under voluntary 

For employer-based programs that struggle with low enrollment 

rates, especially among certain employee subgroups, an automatic 

enrollment strategy may not only increase the total number 

of enrollees but may also decrease some enrollment disparities.

The Diabetes Health Plan included variation in enrollment strategy across employer groups, with some using voluntary enrollment and others using automatic enrollment.

	• Utilizing voluntary enrollment, 35% of eligible employees were enrolled, and annual income ≥$125,000 and black race were associated with much higher rates of enrollment.

	• Utilizing automatic enrollment, 91% of eligible employees were enrolled, and dependents and employees of Hispanic ethnicity were somewhat more likely to be enrolled than other employees.

	• Automatic enrollment strategies can increase overall enrollment and provide participants with the opportunity to “opt out,” and may also decrease some enrollment disparities.

Despite extensive recruitment efforts by health 

plans, state and local governments, and other 

stakeholders, many eligible individuals do not 

voluntarily enroll in health promotion or insurance benefit 

programs designed to improve health outcomes.

are increasingly sponsoring wellness programs as a way to 

possibly decrease costs and increase productivity across a 

use of various approaches, enrollment in wellness programs 

benefits may improve access and outcomes among the subset 

of persons who are enrolled, with limited reach they 

are unlikely to improve the health of the overall targeted 

Many employer health programs use a voluntary enrollment 

approach, which employees must actively join in order 

to be enrolled. However, voluntary program enrollees may 

have different demographic characteristics than the underlying 

population, in terms of gender, age, race/ethnicity, 

income, risk for chronic conditions or disability, and other 

 Voluntary program enrollees may also have different 

clinical characteristics than the underlying population, 

potentially representing either the “worried well” who may 

have less need for services or a sicker subgroup motivated 

Articles

An automatic enrollment strategy for health insurance programs may not only increase the total number of enrollees but may also decrease some enrollment disparities.

Optimizing Enrollment in Employer Health Programs: A Comparison of Enrollment Strategies in the Diabetes Health Plan

 Although research suggests that mail order pharmacy use is associated with greater medication adherence and cardiovascular disease risk factor control, no research has examined the potential impact of mail order pharmacy use on patient safety and utilization.

To compare safety and utilization outcomes in patients using mail order versus local pharmacies.

 Cross-sectional, observational study of 17,217 Kaiser Permanente Northern California adult diabetes patients prescribed new cardiometabolic medications in 2006.

 Multivariate logistic regressions assessed the association between mail order pharmacy use and all-cause and preventable hospitalizations and emergency department (ED) visits; laboratory tests for monitoring persistent medications; and overlapping days of supply of contraindicated medications. Results were stratified by patient age and converted to adjusted predicted percentages.

 Patients aged less than 65 years using mail order had fewer ED visits (33.8% vs 40.2%; 

<.001); preventable ED visits (7.7% vs 9.6%; 

<.01); and serum creatinine laboratory monitoring tests after angiotensin-converting enzyme inhibitor/angiotensin receptor blocker or diuretic initiation (41.2% vs 47.2%; 

 <.01). Among patients aged 65 or more years, mail order users had fewer preventable ED visits (13.4% vs 16.3%; 

 <.01); but slightly more occurrences of overlapping days of supply of contraindicated medications (1.1% vs 0.7%; 

 Mail order pharmacy use is not associated with adverse outcomes in most diabetes patients, and is associated with lower ED use. Interventions to increase mail order pharmacy use should use a patient-centered approach that is sensitive to primary and preventive care access.

Mail order pharmacy use is associated with better medication adherence and cardiovascular disease risk factor control, but its impact on patient safety and utilization is unknown.Among mail order users, we found:

		Diabetes patients aged less than 65 years had fewer emergency department (ED) visits, but fewer monitoring tests after angiotensin-converting enzyme inhibitor/angiotensin receptor blocker or diuretic initiation.

		Diabetes patients aged 65 or more years had fewer preventable ED visits, but slightly more overlapping days of supply of contraindicated medications.

		Interventions to increase mail order pharmacy use may improve outcomes, but should employ patient-centered approaches sesitive to care access.

Mail order pharmacies are widely used to deliver medications in the United States, with up to one-third of chronic illness medications delivered by mail.

 Research suggests a positive association between mail order pharmacy use and greater adherence to diabetes and antihypertensive medications,

 and with better low-density lipoprotein cholesterol control.

 These   favorable outcomes may reflect improved access to medications with mail order pharmacy use,

which may be of particular value to patients with disabilities, time constraints, or limited transportation (

However, there may be unintended consequences to using mail order pharmacies. When medical offices and pharmacies are in the same location, patients who use mail order pharmacies may be less likely to access preventive care services. This may increase the risk of hospitalizations and emergency department (ED) visits that are sensitive to preventive care quality.

Diabetes patients, who often take multiple medications, may be particularly at risk for exposure to contraindicated medications.

Mail order pharmacy users may miss face-to-face consultations with pharmacists and physicians designed to prevent contraindicated medication use. Finally, many medications require laboratory monitoring to reduce the risk of potential adverse drug events,

 and mail order pharmacy use may inadvertently   reduce the likelihood that such tests are ordered by physicians and completed by patients. The purpose of this study was to examine the relationship between mail order pharmacy use and safety and healthcare outcomes in diabetes patients, and whether effects vary by key patient characteristics.

This study was conducted within Kaiser Permanente Northern California (KPNC), an integrated health system providing  comprehensive medical care to more than 3 million members. We selected subjects from the KPNC Diabetes Registry. The registry, established in 1993, is updated annually by adding patients with diabetes identified from automated databases of pharmacy data,laboratory data, hospitalization records, and outpatient diagnoses.

 Study subjects were required to be at least 18 years of age by January 1, 2006, and to have been prescribed a new antiglycemic, antihypertensive, or lipid-lowering medication from January 1 through May 31, 2006. We defined having a new medication as having no history of any medication prescription fill within the same therapeutic drug class within the past 24 months. If patients were prescribed multiple new cardiometabolic medications during this period, the earliest prescription was defined as the index medication. We excluded the small proportion (<5%) of  members who lacked KPNC drug benefits during the study period.

KPNC maintains a mail order pharmacy distribution system in coordination with more than 120 local walk-in KPNC pharmacies located within outpatient clinics and hospitals.

 The first fill of a medication is typically at a local KPNC pharmacy and includes a pharmacist consultation. Patients may then refill existing prescriptions either by mail or at any KPNC local pharmacy. Mail delivery of medications with free shipping can be requested by phone or online. While KPNC members also have the option of filling prescriptions at non-KPNC pharmacies, the cost of these fills is not covered by the patient’s pharmacy benefit. Pharmacists are available via telephone to answer medicationrelated questions regardless of the mode of delivery. There is no proscribed days of supply of pills required for mail delivery; KPNC typically dispenses 100-day supplies through both mail order and local pharmacies. Some patients have a financial incentive to use mail order in the form of a lower copayment for the same number of days of supply.  For this analysis, patients were defined as local KPNC pharmacy users if they never used the mail order pharmacy to fill the new index medication in the 12 months after initiation, and as mail order pharmacy users if they filled the index medication at least once via mail during that time frame.

All-cause hospitalizations, preventable hospitalizations, all-cause ED use, and preventable ED use were defined as no episodes versus 1 or more episodes in the 3 years after the index medication initiation date. Preventable hospitalizations and preventable ED visits were defined using published and validated lists of encounters that are considered sensitive to primary care access and quality (eg, hospitalizations and ED visits for asthma).

Among patients whose index medication was an angiotensin-converting enzyme (ACE) inhibitor, angiotensin receptor blocker (ARB), or diuretic, we examined whether there was an appropriate potassium or serum creatinine lab test within 30 days after the index date.

 These tests monitor for the possibility of electrolyte abnormalities or decreased renal function in patients new to these medications.

Potential for serious medication interaction was defined using a published list of contraindicated medications

 and was operationalized as 2 or more occurrences of overlapping days of supply of contraindicated drugs (eg, anticoagulants and thyroid hormones) for 1 or more days in the 15 months after the index date.

We estimated separate logistic regression models to examine the relationship between mail order pharmacy use and each outcome variable. In each analysis we adjusted for age; sex; race/ethnicity; neighborhood deprivation score calculated with geocoded census socioeconomic data

; number of comorbidities; smoking status; use of nonformulary medications, antidepressants, or insulin; whether the index medication was generic or brand name; length of therapy with the index medication; days of supply of the first refill; distance from the patient’s home to their nearest local KPNC pharmacy; and the patient’s KPNC primary outpatient medical facility (operationalized as a series of dichotomous variables). Detailed definitions of these variables have been published previously.

 Patients with missing covariates (n = 85) were excluded from the analysis. Models included interaction terms for mail order use with age, race/ethnicity, and deprivation score to assess whether different groups of patients might be more vulnerable to potential safety concerns when using the mail order pharmacy. Since the age interaction was the only interaction term that was significant, we present the results as adjusted percentages from models stratified by age less than 65 years versus age 65 years or older.

Omitted (unmeasured) variable bias may be of particular concern when studying the impact of mail order pharmacy use on patient safety outcomes. It is possible that patients who used the mail order pharmacy were different from those who did not in ways we were unable to measure (eg, patient motivation). As a sensitivity analysis, we examined the potential impact of unmeasured variables on the relationship between mail order pharmacy use and our outcomes using methods that test the robustness of effect sizes in the presence of confounding.

All analyses were performed using Stata release 10.1 (StataCorp, College Station, Texas). The study protocol was approved by the KPNC Institutional Review Board.

A total of 17,217 diabetes patients met study eligibility criteria. Of these, 5890 (34.2%) used the mail order pharmacy at least once to refill their new cardiometabolic medication during the study period (

). Mail order pharmacy users were more likely to be white (56.7% vs 37.1%; 

<.001) and to live in census tracts with higher socioeconomic status (

 shows the adjusted predicted percentages of patients experiencing each outcome, stratified by age. After adjustment for demographic, clinical, and census-block characteristics, patients aged less than 65 years using the mail order pharmacy were less likely to have an all-cause ED visit (33.8% vs 40.2%; 

 <.001); less likely to have a preventable ED visit (7.7% vs 9.6%; 

 <.01); and less likely to receive a recommended serum creatinine lab test within 30 days if their index medication was an ACE, ARB, or diuretic (41.2% vs 47.2%; 

 <.01). Among patients aged 65 or more years, those using mail order were less likely to experience a preventable ED visit (13.4% vs 16.3%; 

<.05) and slightly more likely to experience 2 or more occurrences of overlap in days of supply of contraindicated medications (1.1% vs 0.7%; 

Our sensitivity analysis found that our observed effects of mail order use on ED visits, hospitalizations, and overlapping days of supply of contraindicated medications were very robust to the presence of a hypothetical unmeasured confounder; for example, even an unmeasured confounder strongly related to an outcome such as ED use (assuming a relative risk of 2.0) and prevalent in more than 50% of the mail order pharmacy—using population would not change the relationship seen in our findings.

This is the first study to examine the relationship between mail order pharmacy use and patient safety and utilization outcomes. We found that mail order use was not negatively associated with patient safety outcomes overall, suggesting that mail order use does not serve as a barrier to receiving primary and preventive care services for most patients. Patients aged less than 65 years who used the mail order pharmacy did have slightly lower rates of serum creatinine testing with ACE, ARB, or diuretic use (41.4% vs 47.2%; 

 <.01). While there was no statistically significant difference in potassium laboratory testing in this group, or for any lab testing in patients aged 65 or more years, our findings suggest that inperson pharmacy refills may increase opportunities for monitoring of persistent medications in settings where laboratories and pharmacies are co-located.

Our study also found that patients aged 65 or more years had slightly higher rates of 2 or more occurrences of overlapping days of supply of contraindicated medications. While these rates were extremely low for both mail order and local pharmacy users (1.1% and 0.7%, respectively), our analysis adds to the evidence suggesting that older patients using multiple medications may require additional monitoring for adverse events.

 It is important that pharmacies provide increased access to chronic illness medications while also maintaining appropriate systems for preventing contraindicated medication use.

We found mail order pharmacy use was associated with lower overall ED use in patients aged less than 65 years, as well as lower preventable ED use in patients in all age groups. Previous studies have shown mail order use is associated with greater medication adherence

 and low-density lipoprotein cholesterol control.

 Since good medication adherence and cardiovascular disease risk factor control are associated with reduced hospitalizations and better outcomes,

mail order use may lead to lower preventable utilization over time.

Mail order pharmacy services, which are provided at the healthcare delivery system level rather than the physician or practice level, can be considered a structural or system-level intervention for improving access to chronic illness medications.

suggest that using mail order pharmacy services may improve outcomes, our study also suggests that certain patients may be more sensitive to any potential negative impact of mail order pharmacy use on healthcare access. System-level efforts to promote mail order use should preserve patient choice in pharmacy services and maintain a patient-centered approach to meeting an individual’s overall healthcare needs. Future research should continue to address the important issue of how to appropriately target and promote mail order pharmacy services while preserving the benefits of in-person interaction with pharmacists and other care providers.

This study has several limitations. Observational studies cannot entirely control for differences between patients who use mail order pharmacy services and those who do not. While our sensitivity analysis demonstrated that the relationship between mail order pharmacy use and our outcomes would still be observed even in the presence of a powerful, prevalent confounder, future research should leverage randomized controlled trial designs to examine the impact of mail order pharmacy use on intermediate outcomes, healthcare utilization, and patient safety.

We were unable to ascertain pharmacy utilization at non-KPNC pharmacies. However, we limited our analyses to members with a drug benefit, which is only recognized at KPNC pharmacies. This under-ascertainment should be minimal, as previous research suggests that diabetes patients with drug benefits report extremely low use of non-KPNC pharmacies.

This study was conducted in an integrated delivery system where patients access prescription medications differently than in many other settings. Use of “brick and mortar” pharmacies to fill prescriptions is primarily limited to KPNC local pharmacies; KPNC’s mail order and local pharmacies do not differ in the standard days of supply of pills issued; and mail order pharmacy does not require physician enrollment of the prescription. In addition, KPNC local pharmacies are primarily co-located with medical offices and laboratories, which is often not the case in other settings. This co-location may make any relationship between in-person pharmacy use and preventive care services stronger than in cases where patients obtain in-person refills from brick-and-mortar pharmacies that are not integrated with outpatient care. It is therefore possible that the use of mail order pharmacy services in KPNC has a different impact on outcomes than it would have in other settings.

Mail order pharmacy use is not associated with adverse events in most diabetes patients and may be associated with improved health outcomes. System-level interventions to increase mail order pharmacy use may be an important strategy for improving processes and outcomes of care for patients with chronic illness; however, these interventions should use a patient-centered approach that does not increase risk in older patients and others who may be sensitive to primary and preventive care access.

 From Division of Research (JAS, AJK, WTD), Pharmacy Outcomes Research Group (JC), Kaiser Permanente Northern California, Oakland, CA; David Geffen School of Medicine (KD), University of California, Los Angeles; Los Angeles, CA.

This study was funded by the Division of Diabetes Translation, Centers for Disease Control and Prevention, contract CDC U58 DP002641. This activity was also supported by the Health Delivery Systems Center for Diabetes Translational Research (National Institute of Diabetes and Digestive and Kidney Diseases [NIDDK] grant 1P30-DK092924), the Diabetes Study of Northern California (NIDDK grant R01 DK080726), and the Diabetes and Aging Study (NIDDK grant R01 DK0817960); the Harold Amos Medical Faculty Development Program from the Robert Wood Johnson Foundation; and the National Institute on Aging (grant K08 AG033630). The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the funding organizations.

 The authors (JAS, AJK, WTD, JC, OKD) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (JAS, AJK, OKD); acquisition of data (JAS, AJK, WTD); analysis and interpretation of data (JAS, WTD, JC, OKD); drafting of the manuscript (JAS); critical revision of the manuscript for important intellectual content (AJK, JC, OKD); statistical analysis (WTD); obtaining funding (JAS, AJK); administrative, technical, or logistic support (JC); and supervision (JAS).

Julie A. Schmittdiel, PhD, Division of Research, Kaiser Permanente, 2000 Broadway, Oakland, CA 94612. E-mail: Julie.A.Schmittdiel@kp.org1. Medel M; for the Pharmaceutical Strategies Group. PBM 101. http:// www.psgconsults.com/resources/pbm_101.pdf. Published 2011. Accessed June 24, 2011.

2. Federal Trade Commission. Pharmacy Benefit Managers: Ownership of Mail-Order Pharmacies. Washington, DC: Federal Trade Commission; 2005.

3. Duru OK, Schmittdiel J, Dyer W, et al. Mail-order pharmacy use and adherence to diabetes-related medications. Am J Manag Care. 2010;16(1):33-40.

4. Pittman DG, Tao Z, Chen W, Stettin GD. Antihypertensive medication adherence and subsequent health care utilization and cost. Am J Manag Care. 2010;16(8):568-576.

5. 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.

6. Zhang L, Zakharyan A, Stockl KM, Harada AS, Curtis BS, Solow BK. Mail-order pharmacy use and medication adherence among Medicare Part D beneficiaries with diabetes. J Med Econ. 2011;14(5):562-567.

Patients using mail order pharmacies have lower healthcare utilization, but somewhat less laboratory monitoring of persistent medications and slightly higher contraindicated medication use.

Safety and Effectiveness of Mail Order Pharmacy Use in Diabetes

To investigate whether patients who use mail-order  pharmacies were more likely to have good medication adherence than patients who use local pharmacies.

 We conducted cross-sectional analyses of patients from the Kaiser Permanente Northern California (KPNC) diabetes registry who received a new antiglycemic, antihypertensive, or lipid-lowering index medication between January 1, 2006, and May 31, 2006. We defined good adherence as medication availability at least 80% of the time (ie, a continuous measure of medication gaps value of ≤20%) and compared adherence between mail-order users (≥66% of refills by mail) and KPNC local pharmacy users (all refills in person). Adherence was calculated from the initial dispensing through 15 months of follow-up, medication discontinuation, or May 31, 2007, whichever came first. We analyzed the data using multivariate logistic regression models, after determining that unmeasured patient-level factors and self-selection did not significantly bias our analyses.

 A total of 13,922 eligible patients refilled an index medication. Compared with those who used only local KPNC pharmacies, patients who received medications by mail were more likely to have good adherence (84.7% vs 76.9%, 

 <.001). After adjusting for potential confounders, including days’ supply and out-of-pocket costs, mailorder users had better adherence to antiglycemic, antihypertensive, and lipid-lowering medications (

 Compared with patients who obtained medication refills at local pharmacies, patients who received them by mail were more likely to have good adherence. The association between mail-order use and medication adherence should be evaluated in a randomized clinical trial.

In this study of an integrated delivery system, the use of mail-order pharmacies was associated with a higher likelihood of good medication adherence.

Minority patients and patients of low socioeconomic status were less likely to refill their medications by mail. 

Although these findings should be confirmed in other analyses and in a randomized controlled trial, they suggest the intriguing potential of organization-level interventions to improve adherence, particularly among vulnerable populations, by promoting increased mail-order use.

Although recent scientific breakthroughs have produced medications that improve outcomes for chronic conditions, there has not been a corresponding improvement in medication adherence.

 Although better adherence is associated with fewer hospitalizations and lower mortality in patients with diabetes mellitus,

 half of the patients who are prescribed medications for chronic conditions such as diabetes discontinue them after several years. Of patients who continue taking these medications, only 50% to 70% achieve good adherence,

 in part because of various patient-level, providerlevel, and system-level obstacles.

 Patient—system interactions such as medication refills are also important determinants of medication adherence. Successful and timely refills are not a guarantee of medication consumption, but they represent a minimum step necessary to ensure good adherence.

The use of mail-order pharmacies may streamline the refill process because mailed medications eliminate the need for travel to the pharmacy. Home delivery may be particularly beneficial for patients with disabilities, inadequate access to transportation, or time constraints.

 Patients who use mail-order pharmacies purchase more medications than patients using local walk-in pharmacies.

 However, this difference could represent factors other than better adherence such as self-selection into mail order by adherent patients. In addition, most mail-order pharmacies mandate that patients purchase a 90-day medication supply, unlike local pharmacies. Medication purchases by mail-order users may not reflect actual adherence if providers alter the regimen after 30 to 60 days and patients are obligated to waste pills.

 Although one-third of chronic disease prescriptions in the United States are filled by mail,

 there has been little evaluation of patients’ propensity to use mail-order pharmacies or whether the use of mail-order pharmacies is associated with medication adherence.

Among patients with diabetes in an integrated delivery system, we examined (1) the demographic and clinical factors associated with the use of mail-order pharmacies and (2) the likelihood of good adherence to newly prescribed antiglycemic, antihypertensive, and lipid-lowering medications, comparing mail-order pharmacy users with local pharmacy users. We used a data set in which the days’ supply is similar for mail-order and local pharmacies, and we used statistical techniques to assess for selection bias. We hypothesized that patients who used a mail-order pharmacy would be more likely to have good medication adherence over a 12-month period compared with patients who used only local pharmacies.

We conducted this study within Kaiser Permanente Northern California (KPNC), a fully integrated health system that provides comprehensive medical care to more than 3 million members. The KPNC membership includes employed persons and retirees and approximates the general population of northern California racially/ethnically and socioeconomically. The study protocol was approved by the KPNC Institutional Review Board. We selected subjects from the KPNC diabetes registry, established in 1993.

 The registry is updated annually by adding patients with diabetes identified from automated databases of pharmacy data, laboratory data, hospitalization records, and outpatient diagnoses.

Kaiser Permanente Northern California has more than 120 local walk-in pharmacies, located on-site within outpatient clinics and hospital facilities. We identified each patient’s clinic and local pharmacy based on his or her prior utilization or the utilization patterns of other patients in the same zip code.

 Since 1999, KPNC has maintained a mail-order pharmacy distribution system, in coordination with local KPNC pharmacies. After completing a simple enrollment process, patients can obtain medications by mail. Although most new prescriptions are filled in the local pharmacy, KPNC patients also have the option of filling new prescriptions by mail, with telephone access to a pharmacist to answer any medication-related questions. There is no minimum days’ supply required for mail-order delivery at KPNC, which typically dispenses 100-day medication supplies through mail-order and local KPNC pharmacies. Some patients have a financial incentive to use mail order, typically a lower copayment for the same days’ supply obtained by mail versus at a local KPNC pharmacy.

From the diabetes registry, we selected a subset (n = 23,488) of patients who were aged at least 18 years by January 1, 2006; had a pharmacy benefit; and had been prescribed a new diabetes-related medication (no recorded use in the previous 24 months) between January 1, 2006, and May 31, 2006. To most accurately estimate pharmacy utilization, we excluded members who lacked drug benefits during the study period and had little incentive to use KPNC pharmacies (approximately 4% of patients). Diabetes-related medications included antiglycemic, antihypertensive, or lipid-lowering medications. If patients were prescribed multiple new diabetes-related medications during this period, the earliest prescription was defined as the index medication. We estimated adherence for this medication using a cross-sectional study design (described herein). The study window started at the initial dispensing and continued for 15 months, until discontinuation of the medication, or until May 31, 2007, whichever came first. Patients were considered to have discontinued their medication if they did not obtain a refill within 90 days after their existing supply had run out.

To assess adherence to the index medication, we used a pharmacy utilization—based measure of secondary adherence, the continuous measure of medication gaps (CMG).

 The CMG uses refill data to determine the cumulative period for which no medication was available to the patient (gaps), dividing the number of days for which the patient did not have the medication by the number of days in the study window for that participant. The CMG values range from 0% (completely adherent) to 100% (completely nonadherent). We dichotomized CMG values, classifying values of 20% or less as good and greater than 20% as inadequate. This cutoff has been used in previous studies

 that examined the relationship between medication adherence and hospitalizations and mortality rates. Because the CMG cannot be calculated without at least 1 complete refill interval, we excluded 5613 patients who did not fill their prescription at least twice. The CMG cannot be reliably calculated for insulin given the flexible dosing instructions, so we also excluded 658 patients whose index medication was insulin. Patients with oral index medications who were taking insulin concurrently were not excluded.

We defined 2 patient groups for our analytic comparisons: mail-order users and local pharmacy users. Because the proportion of patients using mail-order exclusively was small (<2% of the sample), we considered patients to be mail-order users (n = 2595) if they used this system predominantly (≥66% of refills by mail). Local pharmacy users (n = 11,327) obtained all their index medication refills in person. Patients who used mail-order intermittently (1%-65% of refills [n = 3295]) were excluded from the analysis.

We included several demographic characteristics in our analyses, including race/ethnicity (non-Latino white, African American, Latino, Asian/Pacific Islander, Native American, mixed race/ethnicity, and missing race/ethnicity). As a contextual measure of socioeconomic status, we used a neighborhood deprivation index20 based on principal components analysis of the following 8 census tract variables from the 2000 US Census: percentage of households in poverty, percentage of households receiving public assistance, percentage of female-headed households with children, percentage of households earning less than $30,000 annually, percentage with less than a high school education, percentage in crowded housing (>1 person per room), percentage unemployed, and percentage of men in management or professional occupations. The socioeconomic deprivation score was standardized to a mean (SD) of 0 (1), with negative scores indicating less deprived neighborhoods.

The number of comorbid conditions for each patient was ascertained from the KPNC outpatient clinical records database for the 18 months before the first refill. Patients were classified by their type of insurance (non-Medicare commercial, Medicare Part D with a group insurer, or Medicare Part D from the individual insurance market) and by whether they had a financial incentive to order by mail. We calculated the linear distance from the patient’s home address to his or her local KPNC pharmacy using MapMarker geocoding software (Pitney Bowes MapInfo, Troy, NY). We categorized the days’ supply of the index medication dispensed at the first refill into 1 to 30, 31 to 60, 61 to 90, or more than 90 days. Finally, we measured the duration of therapy for the index medication, defined as the number of days from the first fill through the last fill before discontinuation or the end of the study window.

We estimated separate logistic regression models using Stata version 9.2 (StataCorp LP, College Station, TX) to examine (1) variables associated with mail-order pharmacy use and (2) differences in the probability of good medication adherence between mail-order and local pharmacy users. In each analysis, we included the following as covariates: age; sex; race/ethnicity; neighborhood socioeconomic deprivation score; number of comorbidities; smoking status; use of nonformulary medications, antidepressants, or insulin; insurance type; financial incentive to order by mail; distance from the patient’s home to the local pharmacy; an indicator for whether the index medication was generic or brand name; length of therapy with the index medication; and medical facility dummy variables. To facilitate interpretability of the results, we estimated predicted percentages (95% confidence intervals) for groups of interest, holding all other covariates at their mean value.

Observational studies are susceptible to more bias than experimental ones when studying causal relationships. Omitted (unmeasured) variable bias, or self-selection (eg, differences in intrinsic patient motivation to use mail order), is of particular concern in this study. We used health econometric techniques, specifically a bivariate probit (BVP) model incorporating an instrumental variable, to assess whether our estimates from the logistic regression predicting medication adherence were subject to omitted variable bias. A nonlinear BVP model is more appropriate than a standard linear instrumental variable model when the outcome (eg, medication adherence) and the treatment variable (eg, mail-order pharmacy use) are binary.

We used distance from the patient’s home to his or her local pharmacy as our instrumental variable in the BVP model based on empirical evidence and theoretical considerations. 

 Greater distance was associated with a greater likelihood of mail-order pharmacy use in unadjusted analyses (

 <.001). In addition, distance was not independently associated with medication adherence or other variables such as age, sex, socioeconomic status, and number of comorbidities. The lack of association between our instrument and these observed variables suggests that it is also independent of other correlated but unobserved variables, although this relationship cannot be directly tested. Finally, multiple health services research studies

 have used distance to healthcare facilities as an instrument.

The likelihood ratio test in the BVP model was nonsignificant (

 = .94), and the observed effect size was similar to the logistic regression model. Because the likelihood ratio test did not provide evidence of omitted variable bias, we elected to report results from the more efficient logistic regression analysis rather than the BVP model.

The final analytic sample included 13,922 patients with diabetes who had been recently prescribed an antiglycemic, antihypertensive, or lipid-lowering medication and filled the prescription at least twice (

). Compared with local KPNC pharmacy users, mail-order pharmacy users were more likely to be non-Latino white (61.0% vs 37.1%) and in the least deprived socioeconomic status quartile (27.5% vs 17.8%) (

 <.001 for both). Compared with local pharmacy users, mail-order pharmacy users were more likely to have a financial incentive to fill their prescriptions by mail (49.6% vs 23.0%) and had a greater mean distance from their home to the local pharmacy (8.0 vs 6.7 miles) (

After multivariate adjustment, characteristics associated with mail-order pharmacy use were largely unchanged. Holding all other covariates constant at their mean, the predicted probabilities of mail-order pharmacy use did not differ by patient sex, age, or number of comorbidities (

). However, non-Latino white patients were predicted to have a higher likelihood of mail-order use (24.1%) than African Americans (4.0%), Latinos (5.2%), Asian/Pacific Islanders (8.4%), and patients of mixed race/ethnicity (8.0%) (

<.001 for all). Twenty-one percent of patients living in the least deprived neighborhoods were predicted to refill their medications by mail compared with 12.1%, 10.8%, and 8.4% (

<.001 for all) of patients in progressively more deprived neighborhood quartiles. Patients living at least 4 miles from their local pharmacy were predicted to refill their new medications by mail more often than patients living less than 4 miles from their local pharmacy (16.5% vs 13.3%, 

<.001). Patients with a financial incentive to use the mail-order pharmacy had a substantially greater predicted probability of using mail order compared with patients with no incentive (35.3% vs 10.1%, 

In adjusted models, patients who filled any new prescription by mail were 7.8 percentage points more likely to have good adherence compared with those who used local pharmacies (84.7% vs 76.9%, 

). Separate models for antiglycemic medications (difference of 6.6 percentage points), antihypertensive medications (difference of 7.7 percentage points), and lipid-lowering medications (difference of 9.3 percentage points) (

Our analysis is one of the first studies to examine the characteristics of patients using mail-order pharmacies. It confirmed that patients who received newly prescribed medications through the mail were more likely to have good adherence than patients who obtained them at local KPNC pharmacies. Although there is a paucity of literature in this area, our findings of a 7— to 8–percentage point increase in good adherence associated with mail-order use are consistent with a recent industry non–peer-reviewed study.

 That study compared adherence within a health plan that required members to use a mail-order pharmacy and another plan that required members to use a local pharmacy. The authors found that, compared with the use of a local pharmacy, the use of a mail-order pharmacy was associated with an 8.4—percentage point increase in good adherence to antiglycemic medications, a 7.8–percentage point increase in good adherence to antihypertensive medications, and a 6.8–percentage point increase in good adherence to lipid-lowering medications.

Inferences from observational studies are limited by biases not evident in randomized exposures. Self-selection is of particular concern when studying outcomes such as medication adherence because it may introduce unmeasured differences in patient characteristics. These characteristics could potentially be associated with adherence and result in spurious associations between adherence and pharmacy type. The results of the likelihood ratio test within a BVP model suggest that unmeasured patient self-selection into mail order did not significantly distort our findings. We also controlled for potential confounders not addressed in prior published analyses, including differences in days’ supply and out-of-pocket costs between mail-order and local pharmacies. Although we cannot conclusively determine a causal relationship, the association between mail-order use and medication adherence deserves further investigation with a randomized study design.

Our findings reinforce the importance of system-level determinants of patient behavior. In response to a prevailing attitude of blaming the patient for poor adherence, the World Health Organization

 has developed an adherence model that considers health system and socioeconomic factors, as well as patient-specific and medication-specific factors. Recent qualitative studies

 found that physicians rarely address system-level barriers to medication acquisition when prescribing a new medication. Physicians, nurses, or clinicbased support staff may be able to improve adherence when a new medication is prescribed by ensuring that patients who have efficient access to a reliable mail-order pharmacy system consider using it to obtain their refills.

Our findings also suggest the intriguing potential of larger organization-level interventions to improve adherence by promoting increased mail-order use. Recent review articles

 examining adherence intervention research take a despairing tone, noting that the field is “stuck,” without good options to move forward. Although there have been more than 50 randomized controlled trials of adherence interventions for chronic medications, none have evaluated efforts to minimize system-related barriers to medication acquisition.

 focused on intensive efforts by pharmacists to provide ongoing medication counseling and demonstrated improved adherence and better outcomes. However, such approaches are costly and typically are not compensated by thirdparty payers. In current clinical settings, counseling at the time of medication refills is uncommon.

 Although proven effective, adoption of intensive counseling protocols may require a major expansion of professional responsibilities and represent a significant additional expense for health plans or employers.

Developing effective strategies to increase mail-order pharmacy use could move the field of adherence intervention research in an important new direction. Our findings provide evidence supporting the development and evaluation of interventions to increase mail-order use. Although these interventions would not enhance existing levels of pharmacist counseling, they would potentially be appealing for several other reasons. Assuming similar generic dispensing rates across pharmacy type, mail-order pharmacies seem to be cost saving for a health system after the associated increase in medication use is taken into account.

 Many patients save money when ordering their medications by mail, and efforts to increase mailorder use among interested patients may be well accepted by patients and health systems. Interventions to increase mail-order use would, in many cases, take advantage of existing system infrastructure, further promoting sustainability. Many health systems, including the Department of Veterans Affairs,

 have well-established mail-order pharmacy systems that efficiently fill several million prescriptions per year.

Despite their potential benefits, interventions to increase mail-order pharmacy use would need to be carefully designed and implemented to avoid inadvertently harming patients. Such interventions should not come at the expense of face-to-face pharmacist time needed to clarify dosing schedules or check for adverse drug interactions, including dietary supplements or over-the-counter medications. Research studies that carefully assess the potential benefits and harms of increasing mail-order pharmacy use on a large scale are warranted.

Our study has several limitations. First, it was cross-sectional and cannot conclusively determine a causal relationship. Second, because of sample size limitations, we were unable to compare adherence among patients who use mail-order exclusively versus those who use local pharmacies exclusively. Third, we did not measure medication consumption, although our claims-based measure has been validated against pill counts.

 Fourth, the findings may not generalize to fee-for-service settings with less integrated medication delivery systems or to patients who do not have diabetes. Fifth, we measured adherence to a single medication rather than multiple medications taken simultaneously. Sixth, we measured adherence only up to 15 months after the first medication fill, so our findings may not reflect long-term changes in adherence.

In summary, refilling a newly prescribed diabetes-related medication by mail is associated with a greater likelihood of good adherence compared with refilling at a local pharmacy. Minority patients and patients of low socioeconomic status have particularly low rates of mail-order pharmacy use. Although our findings should be confirmed in a randomized clinical trial, encouraging the use of existing efficient mail-order pharmacies may represent a novel, low-cost, and sustainable approach to improve medication adherence.

: From the Division of General Internal Medicine, Department of Medicine, The David Geffen School of Medicine at UCLA (OKD), University of California, Los Angeles, CA; and the Division of Research (JAS, WTD, MMP, CSU, AJK), and the Pharmacy Outcomes Research Group (JC), Kaiser Permanente Northern California, Oakland, CA.

This study was jointly funded by program announcement 04005 from the Centers for Disease Control and Prevention (CDC) (Division of Diabetes Translation) and the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the CDC or the NIDDK. No authors have financial or other conflicts of interest related to this article and specifically have no ties to local or mail-order pharmacy organizations. Significant contributions to this study were made by members of the Translating Research Into Action for Diabetes (TRIAD) Study Group. Dr Duru had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The funders of this study did not influence the design or conduct of the study; the collection, management, analysis, or interpretation of the data; or the preparation, review, or approval of the manuscript.

The authors (OKD, JAS, WTD, MMP, CSU, JC, AJK) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article. 

Concept and design (OKD, JAS, WTD, MMP, JC, AJK); acquisition of data (JAS, WTD, MMP, CSU, AJK); analysis and interpretation of data (OKD, JAS, WTD, MMP, CU, JC, AJK); drafting of the manuscript (OKD, JAS, WTD); critical revision of the manuscript for important intellectual content (OKD, JAS, WTD, JC, AJK); statistical analysis (JAS, WTD, CU); obtaining funding (AJK); and supervision (OKD).

O. Kenrik Duru, MD, MSHS, Division of General Internal Medicine, Department of Medicine, The David Geffen School of Medicine at UCLA, University of California, Los Angeles, 911 Broxton Plaza, Los Angeles, CA 90024. E-mail: kduru@mednet.ucla.edu.

 Variations in patients’ adherence to medical recommendations: a quantitative review of 50 years of research. 

2. Wens J, Van Casteren V, Vermeire E, Van Royen P, Pas L, Denekens J.

Patients with diabetes were more likely to have good medication adherence if they refilled their medications by mail versus at localpharmacies.

O. Kenrik Duru, MD, MSHS

Articles

Optimizing Enrollment in Employer Health Programs: A Comparison of Enrollment Strategies in the Diabetes Health Plan

Safety and Effectiveness of Mail Order Pharmacy Use in Diabetes

Mail-Order Pharmacy Use and Adherence to Diabetes-Related Medications

Cost Evaluation of a Coordinated Care Management Intervention for Dementia