ORIGINAL INVESTIGATION
Health System Factors and Antihypertensive
Adherence in a Racially and Ethnically Diverse
Cohort of New Users
Alyce S. Adams, PhD; Connie Uratsu, RN; Wendy Dyer, MS; David Magid, MD, MPH;
Patrick O’Connor, MD, MA, MPH; Arne Beck, PhD; Melissa Butler, PharmD, MPH, PhD;
P. Michael Ho, MD, PhD; Julie A. Schmittdiel, PhD
Background: The purpose of this study was to identify
potential health system solutions to suboptimal use of
antihypertensive therapy in a diverse cohort of patients
initiating treatment.
Methods: Using a hypertension registry at Kaiser Permanente Northern California, we conducted a retrospective cohort study of 44 167 adults (age, ⱖ18 years) with
hypertension who were new users of antihypertensive
therapy in 2008. We used multivariate logistic regression analysis to model the relationships between race/
ethnicity, specific health system factors, and early nonpersistence (failing to refill the first prescription within
90 days) and nonadherence (⬍80% of days covered during the 12 months following the start of treatment), respectively, controlling for sociodemographic and clinical risk factors.
Results: More than 30% of patients were early nonpersistent and 1 in 5 were nonadherent to therapy. Nonwhites were more likely to exhibit both types of suboptimal medication-taking behavior compared with whites.
In logistic regression models adjusted for sociodemo-
Author Affiliations: Kaiser
Permanente Division of
Research, Oakland, California
(Drs Adams and Schmittdiel
and Mss Uratsu and Dyer);
Institute for Research, Kaiser
Permanente, Denver, Colorado
(Drs Magid, Beck, and Ho);
HealthPartners Research
Foundation, Minneapolis,
Minnesota (Dr O’Connor);
Kaiser Permanente Center for
Health Research Southeast,
Atlanta, Georgia (Dr Butler);
and Denver Veterans Affairs
Medical Center, Denver
(Dr Ho).
H
graphic, clinical, and health system factors, nonwhite race
was associated with early nonpersistence (black: odds ratio, 1.56 [95% CI, 1.43-1.70]; Asian: 1.40 [1.29-1.51];
Hispanic: 1.46 [1.35-1.57]) and nonadherence (black: 1.55
[1.37-1.77]; Asian: 1.13 [1.00-1.28]; Hispanic: 1.46 [1.311.63]). The likelihood of early nonpersistence varied between Asians and Hispanics by choice of first-line therapy.
In addition, racial and ethnic differences in nonadherence were appreciably attenuated when medication copayment and mail-order pharmacy use were accounted
for in the models.
Conclusions: Racial/ethnic differences in medicationtaking behavior occur early in the course of treatment.
However, health system strategies designed to reduce patient co-payments, ease access to medications, and optimize the choice of initial therapy may be effective tools
in narrowing persistent gaps in the use of these and other
clinically effective therapies.
JAMA Intern Med. 2013;173(1):54-61.
Published online December 10, 2012.
doi:10.1001/2013.jamainternmed.955
EART DISEASE IS THE LEAD-
ing cause of death in the
United States and costs
more than $315 billion
each year in health care
costs and loss in productivity.1,2 Hypertension is a major risk factor for heart disease, and modest reductions in blood pressure have been associated with significant
reductions in the risk of adverse cardiovascular events such as stroke, coronary
heart disease, and death.3 Despite the widespread availability of clinically efficacious medications for treating hypertension, fewer than one-third of patients with
hypertension achieve recommended levels of blood pressure control.4
Some racial and ethnic subgroups may
be at higher risk for inadequate blood pressure control and lower rates of antihypertensive adherence, even in settings with
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
54
somewhat equal access to care.5-9 Suboptimal use of antihypertensives can occur
at any stage in the process, from picking
up the first prescription to making adherence a part of everyday life.4 However, because long-term adherence depends on
persistence with therapy at an early stage
of treatment, early nonpersistence may present a critical opportunity for identifying
practices and policies with the potential
to reduce disparities in cardiovascular
outcomes.
See also page 46
The purpose of this study was to identify potential health system solutions to
suboptimal adherence and persistence with
antihypertensive medications at an early
stage of treatment among a diverse co-
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
Author Affil
Permanente
Research, Oa
(Drs Adams
and Mss Ura
Institute for
Permanente
(Drs Magid,
HealthPartn
Foundation,
Minnesota (
Kaiser Perm
Health Resea
Atlanta, Geo
and Denver
Medical Cen
(Dr Ho).
hort of hypertension patients with equal access to health
insurance. Identifying potentially modifiable health system–level determinants of appropriate medication use
across racial and ethnic groups could inform the development of interventions designed to reduce disparities
in hypertension control.
As a secondary outcome, we examined systolic blood pressure (SBP), a highly reliable measure of blood pressure control across age groups,16-19 measured within 90 days before the
12-month anniversary of the date of therapy start. Good control was defined as 140 mm Hg (⬍130 mm Hg if diabetes mellitus or chronic kidney disease was present).
KEY COVARIATES
METHODS
SETTING AND STUDY POPULATION
This retrospective cohort study was conducted at Kaiser
Permanente Northern California, an integrated health care
delivery system serving more than 3.3 million people. From a
hypertension registry including nearly 1.3 million patients
from 2000 through 2009 identified through a complex algorithm including diagnosis codes and consecutive blood pressure measurements,10 we identified 56 274 adults (age, ⱖ18
years) who were new users of antihypertensive therapy (defined as no evidence of antihypertensive drug dispensing during the previous 8-year period) January 1 through December
31, 2008. We excluded 7818 patients (13.9%) who were not
continuously enrolled and who did not have an active drug
benefit on the date that therapy was started and for at least
250 days following the therapy start date to ensure adequate
follow-up. We allowed for gaps in drug coverage of no more
than 60 days. We also excluded 4288 patients (7.6%) who
were hospitalized at any point during that period. After
excluding 1 additional patient of unknown sex, our final
cohort included 44 167 patients.
DATA SOURCES
The hypertension registry included clinical databases extracted from integrated electronic medical records at Kaiser
Permanente Northern California. Ambulatory pharmacy data
included the date of prescription of medications and refill information. In this closed pharmacy system, patients have a strong
financial incentive to fill prescriptions at a health plan pharmacy, with more than 95% of the members obtaining prescription medications in-house.11,12
DEFINITION OF OUTCOME MEASURES
We identified 3 stages of suboptimal medication-taking behavior among patients receiving new treatment: primary nonadherence, early nonpersistence, and nonadherence. Primary nonadherence was defined as being prescribed a new antihypertensive
medication but failing to fill the prescription within 2 months
of the date it was ordered. Early nonpersistence was defined as
filling the first prescription but failing to refill an antihypertensive medication within 90 days of the date of the first filled
prescription for that medication. Patients who switched medications within the first 90 days of initiating therapy were not
included in the calculation of primary nonadherence or early
nonpersistence.
Nonadherence was defined as not having medication available for 20% or more of days during the 12 months following
initiation of therapy. This measure was calculated only for patients continuously enrolled for at least 12 months after initiation of antihypertensive medication with at least 1 refill
within the first 90 days (ie, excluding patients with early nonpersistence). Continuous gap measures are refill-based measures of nonadherence demonstrating high reliability and validity, particularly when both prescribing and dispensing data
are used.13-15
Race and ethnicity were key covariates in this analysis, and data
were available from administrative sources and self-report. We
classified patients as white (non-Hispanic) (37.0%), black (nonHispanic) (6.9%), Asian (non-Hispanic) (8.8%), Hispanic
(10.1%), and other or unknown (37.2%). Nearly all (97.9%)
patients whose race/ethnicity was categorized as other or unknown had missing data.
We created 3 measures to represent potentially modifiable
system-level determinants of suboptimal medication-taking behavior. First, we identified the level of expected co-payment
for generic medications ($0-$5, $6-$10, and ⬎$10 per prescription) for each patient. More than 95% of all prescribed antihypertensive agents were generic formulations. We also assessed voluntary enrollment in a mail-order pharmacy program,
which included a financial incentive in the form of reduced copayments for approximately 30% of the patients. Use of a mailorder pharmacy was considered a predictor of nonadherence
because enrollment in the program ruled out the possibility of
early nonpersistence (ie, failure to refill the first prescription).
The choice of initial antihypertensive therapy is influenced
by patient and provider characteristics and preferences, as well
as by health system factors, such as clinical guidelines and preferred drug lists. Therefore, we included the medication used
in initiating therapy (diuretics, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, ␤-blockers,
calcium channel blockers, and other agents, including ␣1adrenergic antagonists, ␣2-adrenergic agonists, and peripheral
vasodilators) as a third factor that is potentially modifiable
through health system intervention. Patients initiating therapy
with more than 1 agent (24%) were assigned to the agent listed
first in the pharmacy record.
We further adjusted for factors that may be correlated with
race/ethnicity, including patient age and sex. We also created
categorical variables representing the median household income (⬍$40 000, $40 000-$74 999, or ⱖ$75 000 per year) and
average educational attainment (⬍10%, 10%-19%, 20%-29%,
or ⱖ30% with a bachelor’s degree) for each patient’s US Census 2000 block group of residence.
To control for severity of hypertension, we included the most
recent SBP reading (⬍140, 140-149, 150-159, or ⱖ160 mmHg)
recorded before initiation of antihypertensive treatment, which
was available for 93.4% of the study population. Patientreported smoking status (yes/no) and clinically assessed body
mass index (calculated as weight in kilograms divided by height
in meters squared) were also included.
We also included indicators for competing health care
needs,20 including several medical (cardiovascular disease,
chronic kidney disease, and diabetes mellitus) and mental health
(schizophrenia, bipolar disorder, anxiety, and depression) comorbidities. We used the International Classification of Diseases, Ninth Revision, Clinical Modification21 to identify conditions based on 1 inpatient or 2 outpatient diagnoses observed
between 2000 and the date of initiation of antihypertensive
therapy. Given challenges in identifying depression using medical records,22 we only counted depression diagnoses that occurred during the 12-month period before initiation of antihypertensive medication and used dispensing of select
antidepressant classes (tricyclics, serotonin reuptake inhibi-
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
55
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
tors, and norepinephrine reuptake inhibitors) to identify patients with possible depression. We further controlled for variation in the use of health services by including the number of
medical office visits during the 12 months before starting antihypertensive therapy.
STATISTICAL ANALYSIS
We used logistic regression with time-dependent covariates to
estimate early nonpersistence and nonadherence, respectively, among patients with hypertension who were new users
of antihypertensive therapy. We first estimated the association between race/ethnicity and the primary outcomes, controlling only for age and sex. To evaluate whether health system factors acted as potential mediators of the relationship
between race/ethnicity and medication use, we sequentially
added groups of covariates representing cardiovascular risk factors, socioeconomic status, medical comorbidity, psychiatric
comorbidity, number of medical visits, and finally, the 3 key
health system factors to the model and observed the effect on
the coefficient for race/ethnicity. A relative change in the coefficient of 10% or more was considered strong evidence of a
mediating or explanatory factor.
We also tested for interactions between race/ethnicity and
health system factors to test for possible effect modification using
the log-likelihood statistic. To facilitate the interpretation of
statistically significant interactions, we used the estimates generated by the model with interaction terms to calculate predicted probabilities for each of the main outcomes for important racial/ethnic and medication subgroups. For this calculation,
we set the remaining covariates to match the distribution of these
factors among white patients and then graphically compared
the resulting subgroups of interest. We used multiple imputation to address missing values for body mass index (28.2% missing) and baseline SBP (6.6% missing)23 and compared the model
results with and without imputed values.
Finally, we visually compared changes in SBP over time,
stratifying by race/ethnicity and medication persistence. All statistical analyses were conducted using commercial software (SAS,
version 9.1; SAS Institute Inc).24 This study was approved by
the institutional review board at Kaiser Foundation Research
Institute.
RESULTS
At the time therapy was initiated, SBP was similar across
racial/ethnic subgroups (Table 1). Black and Hispanic
patients were younger, more likely to be obese, and more
likely to have lower household income and educational
attainment compared with other subgroups. Black and
white patients were more likely to be active smokers. Rates
of medical and psychiatric comorbidity were fairly similar by race/ethnicity.
Primary nonadherence (failing to fill a prescription)
was generally low, affecting fewer than 5% of all patients, and was similar across racial/ethnic subgroups (data
not shown). Unadjusted prevalence of early nonpersistence (Table 2) ranged between 11.3% and 42.5% and
was highest for blacks. Unadjusted prevalence of nonadherence (Table 2) ranged between 17.1% and 28.1%
and was also highest for black patients.
The modeling results for early nonpersistence are presented in Table 2. Adjusting only for age and sex, blacks
(odds ratio [OR], 1.59; 95% CI, 1.46-1.73), Asians (1.36;
1.26-1.47), and Hispanics (1.48; 1.37-1.58) all had higher
odds of early nonpersistence compared with whites. Sequential adjustment for cardiovascular risk factors, socioeconomic status, medical and psychiatric comorbidity, frequency of medical visits, and health system factors
did not attenuate the associations between race/
ethnicity and early nonpersistence (Table 2).
Medication co-payment and the choice of antihypertensive therapy were independently associated with early
nonpersistence. A co-payment of $6 to $10 was marginally associated with higher odds of early nonpersistence
relative to a co-payment less than $6 (OR, 1.06; 95% CI,
1.01-1.11). Patients initiating treatment with angiotensin II receptor blockers had a dramatically lower likelihood of early nonpersistence compared with patients
initiating therapy with diuretics (OR, 0.48; 95% CI, 0.400.57). Tests of interactions between race/ethnicity and
these health system factors revealed modification of the
race/ethnicity effect by type of medication (Figure). Specifically, Asians using angiotensin-converting enzyme inhibitors had a higher likelihood of early nonpersistence
(estimated proportion: 38.7%; for interaction, P = .004),
as did Hispanics who initiated therapy with ␤-blockers
(35.0%; P = .06). In contrast, Asians who initiated therapy
with ␤-blockers (24.8%; P = .002) had lower odds of early
nonpersistence with antihypertensive therapy.
Several sociodemographic and clinical factors were associated with early nonpersistence, including younger age,
male sex, smoking, having a body mass index less than
25, baseline SBP between 140 and 149 mm Hg, lower annual income (⬍$40 000), lower educational attainment
(⬍10% bachelor’s degree), diabetes mellitus, and 3 or
more medical visits during the 12 months before initiation of therapy. The results were robust to the inclusion
of imputed values for body mass index and baseline SBP.
Table 2 reports the results of the models predicting
antihypertensive nonadherence. Adjusting for age and sex,
race/ethnicity was associated with nonadherence, with
black (OR, 1.74; 95% CI, 1.53-1.97), Asian (1.20; 1.071.35), and Hispanic (1.67; 1.51-1.86) patients having
higher odds of nonadherence compared with whites.
Medication co-payment, type of medication initiated, and enrollment in a mail-order pharmacy were all
associated with nonadherence. Co-payments of $6 to $10
(OR, 1.19; 95% CI, 1.11-1.28) and $11 or more (1.35;
1.20-1.52) were associated with higher odds of nonadherence during the first year of therapy. Patients initiating therapy with angiotensin II receptor blockers (OR,
1.27; 95% CI, 1.04-1.55) and other less commonly prescribed medications (1.43; 1.18-1.73) had higher odds
of nonadherence compared with patients using diuretics. Enrollment in mail-order pharmacy was associated
with reduced odds of nonadherence (OR, 0.57; 95% CI,
0.53-0.61).
The association between race/ethnicity and nonadherence was attenuated by the inclusion of information about
co-payment status and mail-order pharmacy use. After
inclusion of these covariates, the association between race/
ethnicity and nonadherence was as follows: blacks, OR 1.55
(95% CI, 1.37-1.77); Asians, 1.13 (1.00-1.28); Hispanics,
1.46 (1.31-1.63); and other/unknown, 1.01 (0.93-1.09).
We found no evidence of an interaction between health
system factors and race/ethnicity.
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
56
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
Table 1. Characteristics of Kaiser Permanente Northern California Enrollees
With Newly Initiated Antihypertensive Medications in 2008
No. (%) a
Characteristic
Total
Total
Age, y
⬍50
50-64
65-74
ⱖ75
Sex
Male
Female
Current smoker
Yes
No
BMI
⬍18.5
18.5-24.99
25-29.99
ⱖ30
Missing, No.
Annual household income, $
⬍40 000
40 000-74 999
ⱖ75 000
Missing, No.
Bachelor’s degree, % b
⬍10
10-19
20-29
ⱖ30
Missing, No.
Drug co-payment, $
0-5
6-10
⬎10
Missing, No.
Mail-order pharmacy c
Comorbid conditions
Diabetes mellitus
Prior cardiovascular disease
Chronic kidney disease
Schizophrenia
Bipolar disorder
Anxiety
Depression
Outpatient visits, mean (SD), No. d
SBP, mean (SD), mm Hg e
Drug class
Diuretic
ACE inhibitor
ARB
␤-Blocker
Calcium channel blocker
Other f
44 167
White
Black
Asian
Hispanic
Other/Unknown
16 343 (37.0)
3036 (6.9)
3893 (8.8)
4479 (10.1)
16 416 (37.2)
18 122 (41.0)
18 817 (42.6)
4966 (11.2)
2262 (5.1)
5205 (31.8)
7176 (43.9)
2506 (15.3)
1456 (8.9)
1650 (54.3)
1112 (36.6)
192 (6.3)
82 (2.7)
1681 (43.2)
1574 (40.4)
462 (11.9)
176 (4.5)
2330 (52.0)
1584 (35.4)
402 (9.0)
163 (3.6)
7256 (44.2)
7371 (44.9)
1404 (8.6)
385 (2.4)
22 371 (50.7)
21 796 (49.3)
7870 (48.2)
8473 (51.8)
1247 (41.1)
1789 (58.9)
1590 (40.8)
2303 (59.2)
2034 (45.4)
2445 (54.6)
9630 (58.7)
6786 (41.3)
4653 (10.5)
39 514 (89.5)
2014 (12.3)
14 329 (87.7)
473 (15.6)
2563 (84.4)
275 (7.1)
3618 (92.9)
409 (9.1)
4070 (90.9)
1482 (9.0)
14 934 (91.0)
179 (0.6)
5959 (18.8)
10 893 (34.4)
14 668 (46.3)
12 468
86 (0.7)
2522 (19.4)
4464 (34.4)
5922 (45.6)
3349
⬍10 (0.2)
247 (10.6)
634 (27.3)
1436 (61.8)
714
39 (1.3)
1083 (36.5)
1164 (39.3)
679 (22.9)
928
14 (0.4)
358 (9.9)
1093 (30.2)
2151 (59.5)
863
35 (0.4)
1749 (17.8)
3538 (36.1)
4480 (45.7)
6614
8304 (19.0)
24 487 (55.7)
11 146 (25.4)
230
2553 (15.7)
9206 (56.6)
4517 (27.8)
67
1158 (38.4)
1521 (50.4)
339 (11.2)
18
441 (11.4)
2154 (55.6)
1280 (33.0)
18
1089 (24.5)
2653 (59.6)
708 (15.9)
29
3063 (18.8)
8953 (54.9)
4302 (26.4)
98
9316 (21.3)
14 960 (34.1)
11 973 (27.3)
7955 (18.1)
353
2708 (16.6)
5362 (32.9)
4778 (29.4)
3428 (21.1)
67
1019 (33.8)
1032 (34.2)
681 (22.6)
286 (9.5)
18
505 (13.0)
1234 (31.8)
1270 (32.8)
866 (22.3)
18
1599 (35.9)
1643 (36.9)
860 (19.3)
347 (7.8)
30
3485 (21.4)
5419 (33.2)
4384 (26.9)
3028 (18.6)
100
13 833 (31.6)
24 434 (55.8)
5547 (12.7)
353
9856 (34.4)
5126 (31.6)
8495 (52.4)
2581 (15.9)
141
4704 (42.8)
1346 (44.6)
1413 (46.8)
261 (8.6)
16
305 (18.4)
1303 (33.6)
2107 (54.3)
447 (11.5)
36
626 (26.7)
1348 (30.2)
2568 (57.6)
545 (12.2)
18
511 (20.2)
4710 (28.9)
9851 (60.5)
1713 (10.5)
142
3710 (33.4)
3152 (7.1)
959 (2.2)
909 (2.1)
147 (0.3)
338 (0.8)
2641 (6.0)
1484 (3.4)
5.9 (10.2)
144.3 (17.1)
1239 (7.6)
638 (3.9)
588 (3.6)
69 (0.4)
229 (1.4)
1386 (8.5)
705 (4.3)
8.0 (12.3)
144.0 (17.0)
247 (8.1)
55 (1.8)
49 (1.6)
20 (0.7)
17 (0.6)
194 (6.4)
104 (3.4)
7.1 (11.8)
145.1 (16.3)
348 (8.9)
73 (1.9)
63 (1.6)
14 (0.4)
⬍10 (0.2)
195 (5.0)
102 (2.6)
6.3 (9.4)
142.9 (17.2)
513 (11.5)
92 (2.1)
88 (2.0)
20 (0.4)
27 (0.6)
392 (8.8)
183 (4.1)
7.6 (11.8)
143.5 (16.4)
805 (4.9)
101 (0.6)
121 (0.7)
24 (0.1)
56 (0.3)
474 (2.9)
390 (2.4)
2.9 (5.2)
145.1 (17.5)
12 320 (27.9)
19 013 (43.0)
1258 (2.8)
7635 (17.3)
32 462 (5.6)
1479 (3.3)
4441 (27.2)
6706 (41.0)
365 (2.2)
3132 (19.2)
910 (5.6)
789 (4.8)
1176 (38.7)
1087 (35.8)
55 (1.8)
431 (14.2)
197 (6.5)
90 (3.0)
1098 (28.2)
1552 (39.9)
139 (3.6)
750 (19.3)
227 (5.8)
127 (3.3)
1207 (26.9)
2040 (45.5)
85 (1.9)
763 (17.0)
215 (4.8)
169 (3.8)
4398 (26.8)
7628 (46.5)
614 (3.7)
2559 (15.6)
913 (5.6)
304 (1.9)
Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; BMI, body mass index (calculated as weight in kilograms divided by
height in meters squared); SBP, systolic blood pressure.
a Percentages do not include patients with missing data.
b Number and proportion of patients living in a census block where less than 10%, 10% to 19%, 20% to 29%, or 30% or more of adults have a bachelor’s degree.
c Denominator includes only patients who had at least 1 refill of an antihypertensive medication.
d All outpatient visits during the 12 months before initiation of antihypertensive medication.
e Most recent blood pressure reading prior to initiation of antihypertensive medication.
f Included ␣ -adrenergic antagonists, ␣ -adrenergic agonists, and peripheral vasodilators.
1
2
Results were robust to the inclusion of imputed body
mass index and baseline SBP. After adjustment for all co-
variates, the estimated proportion of patients who were
nonadherent by race/ethnicity was as follows: whites,
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
57
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
Table 2. Log Odds of Nonpersistence and Nonadherence Among Kaiser Permanente Northern California Members
With Newly Initiated Antihypertensive Medications in 2008
Early Nonpersistence, OR (95% CI)
Characteristic
Race
White
Black
Asian
Hispanic
Other/missing
Age, y
⬍50
50-64
65-74
ⱖ75
Sex
Male
Female
Nonsmoker
Smoker
BMI
⬍24.99
Overweight (25-29.99)
Obese (ⱖ30)
Missing
SBP, mm Hg
⬍140
140-149
150-159
ⱖ160
Household income, $
⬍40 000
40 000-74 999
ⱖ75 000
Bachelor’s degree, % a
⬍10
10-19
20-29
ⱖ30
No diabetes mellitus
Diabetes mellitus
No CKD
CKD
No prior CVD
Prior CVD
No concurrent depression
Concurrent depression
No schizophrenia
Ever schizophrenia
No bipolar disorder
Ever bipolar disorder
No anxiety
Ever anxiety
Nonadherence, OR (95% CI)
Unadjusted
Proportion, %
Age- and
Sex-Adjusted
Fully
Adjusted
Unadjusted
Proportion, %
Age- and
Sex-Adjusted
Fully
Adjusted
30.6
42.5
38.1
41.1
11.3
1 [Reference]
1.59 (1.46-1.73)
1.36 (1.26-1.47)
1.48 (1.37-1.58)
0.97 (0.92-1.02)
1 [Reference]
1.56 (1.43-1.70)
1.40 (1.29-1.51)
1.46 (1.35-1.57)
1.07 (1.01-1.13)
17.1
28.1
20.7
27.3
19.2
1 [Reference]
1.74 (1.53-1.97)
1.20 (1.07-1.35)
1.67 (1.51-1.86)
1.06 (0.99-1.15)
1 [Reference]
1.55 (1.37-1.77)
1.13 (1.00-1.28)
1.46 (1.31-1.63)
1.01 (0.93-1.09)
39.1
28.7
28.1
31.1
1 [Reference]
0.66 (0.63-0.69)
0.64 (0.59-0.69)
0.75 (0.68-0.83)
1 [Reference]
0.67 (0.64-0.70)
0.65 (0.60-0.71)
0.75 (0.67-0.84)
24.8
17.5
15.2
13.4
1 [Reference]
0.64 (0.60-0.69)
0.56 (0.50-0.62)
0.50 (0.42-0.59)
1 [Reference]
0.66 (0.62-0.71)
0.54 (0.48-0.61)
0.48 (0.40-0.57)
33.1
32.8
32.3
38.8
1 [Reference]
0.96 (0.92-1.00)
20.4
19.1
19.6
21.0
1 [Reference]
0.91 (0.85-0.97)
...
1 [Reference]
0.94 (0.90-0.98)
1 [Reference]
1.16 (1.09-1.24)
...
1 [Reference]
0.95 (0.89-1.01)
1 [Reference]
1.02 (0.92-1.13)
37.3
35.7
36.2
24.8
...
...
...
...
1 [Reference]
0.90 (0.84-0.96)
0.84 (0.78-0.90)
0.66 (0.60-0.72)
16.9
17.8
21.1
21.0
...
...
...
...
1 [Reference]
0.99 (0.89-1.11)
1.08 (0.97-1.20)
1.10 (0.97-1.25)
33.5
36.4
33.3
32.2
...
...
...
...
1 [Reference]
1.06 (1.01-1.12)
0.96 (0.90-1.02)
0.99 (0.93-1.06)
20.2
19.7
19.0
18.5
...
...
...
...
1 [Reference]
0.98 (0.90-1.06)
0.94 (0.86-1.03)
0.92 (0.84-1.01)
35.9
33.0
30.9
...
...
...
1 [Reference]
0.92 (0.87-0.98)
0.91 (0.84-0.99)
22.4
20.0
17.1
...
...
...
1 [Reference]
0.99 (0.91-1.09)
0.95 (0.84-1.07)
36.0
33.4
31.9
30.5
32.6
37.9
33.0
33.6
33.0
31.8
32.8
38.4
33.0
32.4
33.0
38.1
32.6
38.9
...
...
...
...
...
...
...
...
...
...
1 [Reference]
0.96 (0.90-1.02)
0.92 (0.86-0.99)
0.91 (0.83-0.99)
1 [Reference]
1.11 (1.02-1.20)
1 [Reference]
1.01 (0.87-1.18)
1 [Reference]
0.96 (0.82-1.11)
1 [Reference]
1.09 (0.97-1.22)
1 [Reference]
0.76 (0.52-1.10)
1 [Reference]
1.05 (0.83-1.34)
1 [Reference]
1.10 (1.00-1.20)
22.8
20.7
18.3
16.7
19.7
20.1
19.8
18.2
19.8
16.8
19.7
20.6
19.8
14.6
19.7
20.5
19.7
21.0
...
...
...
...
...
...
...
...
...
...
...
...
...
...
1 [Reference]
1.00 (0.91-1.09)
0.95 (0.85-1.05)
0.90 (0.80-1.02)
1 [Reference]
1.00 (0.88-1.13)
1 [Reference]
1.18 (0.93-1.48)
1 [Reference]
1.06 (0.85-1.32)
1 [Reference]
1.07 (0.90-1.28)
1 [Reference]
0.62 (0.34-1.14)
1 [Reference]
1.03 (0.71-1.50)
1 [Reference]
1.05 (0.91-1.20)
...
...
...
...
...
...
(continued)
16.7%; blacks, 28.0%; Asians, 20.3%; Hispanics, 26.9%;
and other/unknown, 18.9%.
Twelve months after therapy was started, blood pressure control improved for all patients. Blood pressure in
fewer than 20% of patients was poorly controlled compared with in more than 60% of patients at baseline. Blacks
who exhibited nonpersistence with antihypertensive
agents had the highest proportion (28.2%) with uncontrolled blood pressure at the end of follow-up. In con-
trast, Asians who were adherent to therapy had the lowest proportion with uncontrolled blood pressure after 12
months (14.1%) (data not shown).
COMMENT
In this setting, we found higher rates of blood pressure control and lower rates of primary nonadherence compared
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
58
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
Table 2. Log Odds of Nonpersistence and Nonadherence Among Kaiser Permanente Northern California Members
With Newly Initiated Antihypertensive Medications in 2008 (continued)
Early Nonpersistence, OR (95% CI)
Characteristic
No. of medical visits
0
1-2
3-5
6-15
Medication co-payment, $
⬍6
6-10
ⱖ11
Drug class
Diuretic
ACE inhibitor
ARB
␤-Blocker
Calcium channel blocker
Other b
No mail-order pharmacy
Mail-order pharmacy
Nonadherence, OR (95% CI)
Unadjusted
Proportion, %
Age- and
Sex-Adjusted
Fully
Adjusted
Unadjusted
Proportion, %
Age- and
Sex-Adjusted
Fully
Adjusted
27.7
34.6
36.3
40.9
...
...
...
1 [Reference]
1.03 (0.96-1.09)
1.08 (1.01-1.17)
1.29 (1.18-1.41)
20.6
19.2
18.6
20.0
...
...
...
1 [Reference]
0.97 (0.89-1.07)
0.95 (0.85-1.06)
1.00 (0.87-1.15)
32.9
33.9
29.1
...
...
...
1 [Reference]
1.06 (1.01-1.11)
0.96 (0.88-1.04)
18.3
20.8
18.7
...
...
...
1 [Reference]
1.19 (1.11-1.28)
1.35 (1.20-1.52)
36.7
32.8
17.3
29.8
26.0
45.1
NA c
NA
...
...
...
...
...
...
...
...
1 [Reference]
0.88 (0.83-0.93)
0.48 (0.40-0.57)
0.78 (0.73-0.83)
0.66 (0.60-0.74)
1.49 (1.33-1.69)
NA
NA
19.1
19.6
22.7
20.1
19.9
22.0
23.0
13.5
...
...
...
...
...
...
...
...
1 [Reference]
1.03 (0.95-1.11)
1.27 (1.04-1.55)
1.09 (0.99-1.20)
1.09 (0.95-1.26)
1.43 (1.18-1.73)
1 [Reference]
0.57 (0.53-0.61)
Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; BMI, body mass index (calculated as weight in kilograms divided by
height in meters squared); CKD, chronic kidney disease; CVD, cardiovascular disease; ellipsis, variable was not included in the model; NA, not applicable;
OR, odds ratio; SBP, systolic blood pressure.
a Refers to patients living in a census block where less than 10%, 10% to 19%, 20% to 29%, or 30% or more of adults have a bachelor’s degree.
b Included ␣ -adrenergic antagonists, ␣ -adrenergic agonists, and peripheral vasodilators.
1
2
c By definition, patients who use mail-order pharmacy received at least 2 refills and therefore could not achieve the outcome of early nonpersistence.
with previous studies.25,26 Good blood pressure control in
this setting has been reported27 and may be related to early
detection, as well as to payment mechanisms that financially reward physician groups based in part on the proportion of patients with good blood pressure control.
Our finding of racial/ethnic differences in both early
nonpersistence and nonadherence is consistent with
prior evidence.8,9,25,26 However, the interaction between
race/ethnicity and medication type as predictors of
early nonpersistence is novel and may reflect unmeasured clinical factors, contraindications, or treatment
preferences that vary across racial/ethnic groups.28-33
Whether variation in early nonpersistence can be reduced through the use of decision tools and other strategies designed to better match patients with therapy deserves further exploration.
Our finding regarding the importance of mail-order
pharmacy use as a potential policy-level mechanism for
improving adherence is consistent with previous studies34-38 in this and other settings. This study provides additional evidence that exposure to a mail-order pharmacy across racial and ethnic subgroups, rather than
differential response to this practice, contributes to disparities in adherence.35 However, all patients using a mailorder pharmacy in this setting also had access to telephone-based counseling from pharmacists, and
approximately one-third also received a financial incentive (ie, lower co-payments) for participating in the program.35 Moreover, integrated electronic medical records may facilitate patient monitoring and continuity of
care for patients receiving medications by mail. Research is needed to explore the specific mechanisms by
which use of a mail-order pharmacy may influence adherence across diverse care settings.
This study has limitations that deserve consideration.
First, we could not control for unmeasured physiological,
behavioral, and psychosocial factors that may explain some
of the observed relationships between race/ethnicity and
behavior. Also, we may have misclassified some patients
because of inaccurate or missing race/ethnicity data and do
not have sufficient information to hypothesize about differences among patients with missing demographic data.
In addition, although the use of pharmacy records to estimate adherence is well supported,13-15 we did not directly observe patient behavior. Therefore, to the extent that
there is greater variation in actual vs estimated adherence,
our findings may be biased. It is possible that we misclassified some patients as having hypertension; however, we
believe that this possibility was reduced through the application of a complex algorithm to identify patients with
hypertension to reduce any potential bias or noise relating to misdiagnosis.10
Evidence from recent disparities-focused interventions in cardiovascular disease indicates that multifaceted interventions to address barriers to self-management may be particularly effective but provide little
guidance as to which barriers are most amenable to
change through health system intervention.39-44 Our
findings suggest that health system factors have the potential to reduce racial and ethnic differences in both
early persistence with and ongoing adherence to antihypertensive therapy. Medication cost and ease of access may be universally important modifiable determinants of ongoing adherence, suggesting that medication
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
59
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
Whites
50
Asians
Hispanics
Patients, %
40
30
20
10
0
Diuretics
ACE Inhibitors
β-Blockers
Drug Class
Figure. Adjusted proportion of patients with early nonpersistence by
antihypertensive drug class and race/ethnicity. ACE indicates
angiotensin-converting enzyme.
Funding/Support: This study was funded by
3U19HL091179-04S1 (National Heart, Lung, and Blood
Institute) and the National Institute for Mental Health
as a supplement to the Health Maintenance Organization Research Network Cardiovascular Disease Network. Additional support for Drs Adams, O’Connor, and
Schmittdiel was provided by P30DK092924 (Health Delivery Systems Center for Diabetes Translational Research), funded by the National Institute for Diabetes and
Digestive and Kidney Diseases.
Role of the Sponsor: The funders had no role in the design or conduct of the study; in data collection, management, analysis, or interpretation; or in the preparation,
review, or approval of this manuscript.
Additional Contributions: Alan Go, MD, commented on
an earlier version of this manuscript, and Karen R.
Hansen, BA, assisted with the preparation of this manuscript for submission.
REFERENCES
cost assistance may be a critical component of interventions to reduce cardiovascular-related disparities. These
findings suggest that the strategies implemented to reduce disparities in medication use should be tailored to
the stage of treatment, recognizing that the relative importance of health system factors is likely to change as
patients move from initiation of therapy to making adherence a part of their everyday lives.
We found evidence that health system factors may play
important roles as mediators and modifiers of racial/
ethnic differences in medication-taking behavior. Unlike socioeconomic and psychosocial determinants that
can be difficult to change, medication choice, copayment, and access are potentially modifiable through
system-level intervention and have the potential to reduce nonadherence, a well-known and especially challenging aspect of hypertension management in highrisk populations.
Accepted for Publication: August 19, 2012.
Published Online: December 10, 2012. doi:10.1001/2013
.jamainternmed.955
Correspondence: Alyce S. Adams, PhD, Kaiser Permanente Division of Research, 2000 Broadway, Oakland, CA
94612 ([email protected]).
Author Contributions: Dr Schmittdiel 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. Study concept and design: Adams, Beck, and
Schmittdiel. Acquisition of data: Uratsu, Dyer, Magid, and
Schmittdiel. Analysis and interpretation of data: Adams,
Uratsu, Dyer, O’Connor, Beck, Butler, Ho, and
Schmittdiel. Drafting of the manuscript: Adams and Beck.
Critical revision of the manuscript for important intellectual content: Adams, Uratsu, Dyer, Magid, O’Connor, Beck,
Butler, Ho, and Schmittdiel. Statistical analysis: Uratsu,
Dyer, and Schmittdiel. Obtained funding: Magid, Beck, and
Schmittdiel. Administrative, technical, and material support: Uratsu, Beck, and Schmittdiel. Study supervision:
Adams and Schmittdiel.
Conflict of Interest Disclosures: None reported.
1. Lloyd-Jones D, Adams RJ, Brown TM, et al; American Heart Association Statistics
Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics—
2010 update: a report from the American Heart Association [published corrections
appear in Circulation. 2011;124(16):e425 and Circulation. 2010;121(12):e260].
Circulation. 2010;121(7):e46-e215. doi:10.1161/CIRCULATIONAHA.109.192667.
2. Qureshi AI, Suri MF, Kirmani JF, Divani AA. Prevalence and trends of prehypertension and hypertension in United States: National Health and Nutrition Examination Surveys 1976 to 2000. Med Sci Monit. 2005;11(9):CR403-CR409.
3. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and control of
hypertension in the United States, 1988-2000. JAMA. 2003;290(2):199-206.
4. Krousel-Wood M, Thomas S, Muntner P, Morisky D. Medication adherence: a
key factor in achieving blood pressure control and good clinical outcomes in hypertensive patients. Curr Opin Cardiol. 2004;19(4):357-362.
5. Cooper R, Cutler J, Desvigne-Nickens P, et al. Trends and disparities in coronary heart disease, stroke, and other cardiovascular diseases in the United States:
findings of the National Conference on Cardiovascular Disease Prevention.
Circulation. 2000;102(25):3137-3147.
6. Fiscella K, Holt K. Racial disparity in hypertension control: tallying the death toll.
Ann Fam Med. 2008;6(6):497-502.
7. Adams AS, Soumerai SB, Ross-Degnan D. Use of antihypertensive drugs by Medicare enrollees: does type of drug coverage matter? Health Aff (Millwood). 2001;
20(1):276-286.
8. Bagchi AD, Esposito D, Kim M, Verdier J, Bencio D. Utilization of, and adherence
to, drug therapy among Medicaid beneficiaries with congestive heart failure. Clin
Ther. 2007;29(8):1771-1783.
9. Batson B, Belletti D, Wogen J. Effect of African American race on hypertension
management: a real-world observational study among 28 US physician practices.
Ethn Dis. 2010;20(4):409-415.
10. Go AS, Magid DJ, Wells B, et al. The Cardiovascular Research Network: a new
paradigm for cardiovascular quality and outcomes research. Circ Cardiovasc Qual
Outcomes. 2008;1(2):138-147.
11. Selby JV, Lee J, Swain BE, et al. Trends in time to confirmation and recognition
of new-onset hypertension, 2002-2006. Hypertension. 2010;56(4):605-611.
12. Saunders KW, Davis RL, Stergachis A. Group health cooperative. In: Strom BL,
ed. Pharmacoepidemiology. 4th ed. New York, NY: John Wiley & Sons Inc; 2005:
223-239.
13. Cramer JA, Roy A, Burrell A, et al. Medication compliance and persistence: terminology and definitions. Value Health. 2008;11(1):44-47.
14. Hansen RA, Kim MM, Song L, Tu W, Wu J, Murray MD. Comparison of methods
to assess medication adherence and classify nonadherence. Ann Pharmacother.
2009;43(3):413-422.
15. Karter AJ, Parker MM, Moffet HH, Ahmed AT, Schmittdiel JA, Selby JV. New prescription medication gaps: a comprehensive measure of adherence to new
prescriptions. Health Serv Res. 2009;44(5, pt 1):1640-1661.
16. Franklin SS, Gustin W IV, Wong ND, et al. Hemodynamic patterns of age-related
changes in blood pressure: the Framingham Heart Study. Circulation. 1997;
96(1):308-315.
17. Wilking SV, Belanger A, Kannel WB, D’Agostino RB, Steel K. Determinants of
isolated systolic hypertension. JAMA. 1988;260(23):3451-3455.
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
60
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017
18. Sagie A, Larson MG, Levy D. The natural history of borderline isolated systolic
hypertension. N Engl J Med. 1993;329(26):1912-1917.
19. Tate RB, Manfreda J, Krahn AD, Cuddy TE. Tracking of blood pressure over a
40-year period in the University of Manitoba Follow-up Study, 1948-1988. Am
J Epidemiol. 1995;142(9):946-954.
20. Redelmeier DA, Tan SH, Booth GL. The treatment of unrelated disorders in patients with chronic medical diseases. N Engl J Med. 1998;338(21):1516-1520.
21. US Department of Health and Human Services; Centers for Disease Control and
Prevention; Centers for Medicare and Medicaid Services. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). 2005-10.
http://hdl.handle.net/1902.29/CD-0177. Accessed June 12, 2008.
22. Trinh NH, Youn SJ, Sousa J, et al. Using electronic medical records to determine
the diagnosis of clinical depression. Int J Med Inform. 2011;80(7):533-540.
23. Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338:b2393.
doi:10.1136/bmj.b2393.
24. SAS Institute Inc. SAS OnlineDoc 9.1. 2002-2003. http://support.sas.com/91doc
/docMainpage.jsp. Accessed June 19, 2012.
25. Shah NR, Hirsch AG, Zacker C, et al. Predictors of first-fill adherence for patients with hypertension. Am J Hypertens. 2009;22(4):392-396.
26. Perreault S, Lamarre D, Blais L, et al. Persistence with treatment in newly treated
middle-aged patients with essential hypertension. Ann Pharmacother. 2005;
39(9):1401-1408.
27. Lester H, Schmittdiel J, Selby J, et al. The impact of removing financial incentives from clinical quality indicators: longitudinal analysis of four Kaiser Permanente indicators. BMJ. 2010;340:c1898. doi:10.1136/bmj.c1898.
28. Horne R, Weinman J. Patients’ beliefs about prescribed medicines and their role
in adherence to treatment in chronic physical illness. J Psychosom Res. 1999;
47(6):555-567.
29. Huang ES, Brown SE, Thakur N, et al. Racial/ethnic differences in concerns about
current and future medications among patients with type 2 diabetes. Diabetes
Care. 2009;32(2):311-316.
30. Schüz B, Wurm S, Ziegelmann JP, Warner LM, Tesch-Römer C, Schwarzer R.
Changes in functional health, changes in medication beliefs, and medication
adherence. Health Psychol. 2011;30(1):31-39.
31. Lehane E, McCarthy G. Intentional and unintentional medication nonadherence: a comprehensive framework for clinical research and practice? a discussion paper. Int J Nurs Stud. 2007;44(8):1468-1477.
32. Clifford S, Barber N, Horne R. Understanding different beliefs held by adherers,
unintentional nonadherers, and intentional nonadherers: application of the Necessity-Concerns Framework. J Psychosom Res. 2008;64(1):41-46.
33. Schnittker J. Misgivings of medicine? African Americans’ skepticism of psychiatric medication. J Health Soc Behav. 2003;44(4):506-524.
34. Duru OK, Schmittdiel JA, Dyer WT, et al. Mail-order pharmacy use and adherence to diabetes-related medications. Am J Manag Care. 2010;16(1):33-40.
35. 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.
36. 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.
37. Pittman DG, Tao Z, Chen W, Stettin GD. Antihypertensive medication adherence
and subsequent healthcare utilization and costs. Am J Manag Care. 2010;16
(8):568-576.
38. 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.
39. Svarstad BL, Kotchen JM, Shireman TI, et al. The Team Education and Adherence Monitoring (TEAM) trial: pharmacy interventions to improve hypertension
control in blacks. Circ Cardiovasc Qual Outcomes. 2009;2(3):264-271.
40. Ogedegbe G, Schoenthaler A, Richardson T, et al. An RCT of the effect of motivational interviewing on medication adherence in hypertensive African Americans: rationale and design. Contemp Clin Trials. 2007;28(2):169-181.
41. Cooper LA, Roter DL, Bone LR, et al. A randomized controlled trial of interventions to enhance patient-physician partnership, patient adherence and high blood
pressure control among ethnic minorities and poor persons: study protocol
NCT00123045. Implement Sci. 2009;4:7. doi:10.1186/1748-5908-4-7.
42. Cutrona SL, Choudhry NK, Fischer MA, et al. Modes of delivery for interventions
to improve cardiovascular medication adherence. Am J Manag Care. 2010;
16(12):929-942.
43. Crook ED, Bryan NB, Hanks R, et al. A review of interventions to reduce health
disparities in cardiovascular disease in African Americans. Ethn Dis. 2009;
19(2):204-208.
44. Chin MH, Walters AE, Cook SC, Huang ES. Interventions to reduce racial and ethnic
disparities in health care. Med Care Res Rev. 2007;64(5)(suppl):7S-28S.
JAMA INTERN MED/ VOL 173 (NO. 1), JAN 14, 2013
61
WWW.JAMAINTERNALMED.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/intemed/926235/ on 06/14/2017