Coping with Prescription Medication Costs: a Cross-sectional Look at Strategies Used and Associations with the Physical and Psychosocial Health of Individuals with Arthritis
Prescription medication costs increase financial burden, often leading individuals to engage in intentional nonadherence. Little is known about what specific medication cost-coping strategies individuals with arthritis employ. The purposes of this study are (1) to identify characteristics of individuals with arthritis who self-report prescription medication cost-coping strategies and (2) to examine the association between medication cost-coping strategies and health status. Seven hundred twenty-nine people self-reporting arthritis and prescription medication use completed a telephone survey. Adjusted regression models examined medication cost-coping strategies and five health status outcomes. Participants reported engaging in cost-coping strategies due to medication costs. Those borrowing money had worse psychosocial health and greater disability; those with increasing credit card debt reported worse physical functioning, self-rated health, and greater helplessness. Medication underuse was associated with worse psychosocial health, greater disability, and depressive symptoms. Individuals with arthritis use multiple strategies to cope with medication costs, and these strategies are associated with adverse physical and psychosocial health status.
Coping with Prescription Medication Costs: a Cross-sectional
Look at Strategies Used and Associations with the Physical
and Psychosocial Health of Individuals with Arthritis
Kathryn Remmes Martin, PhD MPH &
Jack Shreffler, PhD &Britta Schoster, MPH &
Leigh F. Callahan, PhD
#The Society of Behavioral Medicine (outside the USA) 2012
Background Prescription medication costs increase finan-
cial burden, often leading individuals to engage in inten-
tional nonadherence. Little is known about what specific
medication cost-coping strategies individuals with arthritis
Purpose The purposes of this study are (1) to identify
characteristics of individuals with arthritis who self-report
prescription medication cost-coping strategies and (2) to
examine the association between medication cost-coping
strategies and health status.
Methods Seven hundred twenty-nine people self-reporting
arthritis and prescription medication use completed a tele-
phone survey. Adjusted regression models examined medi-
cation cost-coping strategies and five health status
Results Participants reported engaging in cost-coping strat-
egies due to medication costs. Those borrowing money had
worse psychosocial health and greater disability; those with
increasing credit card debt reported worse physical func-
tioning, self-rated health, and greater helplessness. Medica-
tion underuse was associated with worse psychosocial
health, greater disability, and depressive symptoms.
Conclusion Individuals with arthritis use multiple strategies
to cope with medication costs, and these strategies are
associated with adverse physical and psychosocial health
Keywords Medication cost .Medication underuse .Coping
strategies .Health status outcomes .Arthritis
Current prevalence estimates suggest that nearly 50 million
American adults self-report some form of doctor-diagnosed
arthritis, and these numbers are projected to only increase in
the future . Arthritis is not only responsible for lower self-
reported quality of life, increased activity limitation, and
greater disability in the USA [1–4], arthritis also contributes
to increasingly high direct and indirect medical-related costs
[5,6]. At a national level, total aggregate medical expendi-
tures (i.e., inpatient, ambulatory, prescription medications,
home health, emergency room visits) for individuals with
arthritis and other rheumatic conditions (e.g., osteoarthritis,
rheumatoid arthritis, lupus, fibromyalgia) totaled 353 billion
US dollars in 2005—up from 252 billion US dollars in 1997
, and it is estimated that earning losses attributable to
arthritis and other rheumatic conditions were 47 billion US
dollars in 2003 .
At an individual level, high financial costs of medical
care, including prescription medication costs, may present
an additional challenge for adults with arthritis. A national
probability sample of older (70+) US community-dwelling
adults found that those with arthritis, rheumatism, or joint
replacement used more health care services, had greater
health care costs, and higher total out-of-pocket costs, which
K. R. Martin (*)
Laboratory of Epidemiology, Demography, and Biometry,
National Institute on Aging,
Gateway Building, Suite 3C-309, 7201 Wisconsin Avenue,
Bethesda, MD 20892-9205, USA
K. R. Martin :J. Shreffler :B. Schoster :L. F. Callahan
Thurston Arthritis Research Center,
The University of North Carolina at Chapel Hill,
3300 Thurston Bldg; CB 7280,
Chapel Hill, NC 27599-7280, USA
ann. behav. med.
was due in part to those with arthritis more often paying for
prescription drugs not completely covered by health insur-
ance, when compared to those not reporting arthritis, rheu-
matism, or joint replacement . Though general population
estimates vary, it has been suggested that between 20 and
30 % of adults in the USA take medications for any form of
arthritis (e.g., osteoarthritis, rheumatoid arthritis, lupus, fi-
bromyalgia) [9,10]. A greater number of individuals aged
65+ are taking one or more medications (i.e., prescription
medications, over the counter drugs, vitamins, or herbal
supplements) to manage their health , and out-of-
pocket expenses for prescription medicines have consider-
ably increased in this age group . Given that arthritis is
often comorbid with other conditions that require prescrip-
tion medications for disease management (e.g., diabetes and
cardiovascular disease), it stands to reason that the number
of prescription medications being filled has increased, as is
the mean prescription medication expenditures per person
among adults with arthritis, and among adults with arthritis
and 1+ comorbid conditions .
Individuals with arthritis and other chronic conditions
(e.g., heart disease, diabetes) and/or who are older may be
at greater risk for medication underuse either because they
are either underinsured, despite having medical insurance
coverage (e.g., Medicare), or without prescription medica-
tion insurance coverage entirely [9,13]. With many Amer-
icans experiencing rising out-of-pocket health care costs
, individuals must make decisions about the relative
importance of their medical care in relation to their general
survival, as well as decide if and how they will modify their
lifestyles or spending habits in order to afford the medical
care and prescription medications required for good health.
A great deal of popular media focuses attention on the
growing concern of high-cost prescription drugs with
articles that go beyond national statistics and trends to
highlight the personal stories of individuals who struggle
to afford needed medical care [15–17]. Indeed, to cope with
the financial burdens associated with prescription medica-
tion costs, individuals may intentionally become nonadher-
ent by choosing not to take medication as prescribed [18,
19]. Individuals may also self-modulate and take less med-
ication than prescribed (i.e., medication underuse), or make
choices about how they use their available money and credit
to purchase goods and services so that they can afford their
prescription medications (i.e., medication cost-coping strat-
Cost-related medication nonadherence has been
researched in the chronic disease (e.g., cardiovascular, dia-
betes) [21,22] and aging literature . In addition, several
studies have examined cost-related prescription medication
underuse in the general population  and in subgroup
analyses of various chronic conditions, including arthritis
[9,10,23,24]. These studies suggest that cost-related
medication nonadherence is a growing public health prob-
lem, especially given that it has been associated with greater
number of hospitalizations, worse self-reported health, and
poorer physical and mental health outcomes [9,24–28]. In
addition, a recent systematic review has demonstrated that
medication nonadherence among individuals with various
rheumatic conditions is a problem worthy of continued
attention . While prior research examining compliance
with medications among individuals with rheumatic condi-
tions has identified cost and out-of-pocket expenses as a key
factorin treatment adherence [30,31], it appears that no
study has yet explored what specific medication cost-coping
strategies are employed by individuals with arthritis. Given
this gap in the literature, our paper aims to (1) identify the
characteristics of individuals with arthritis who self-report
prescription medication cost-coping strategies and (2) ex-
amine the association between coping strategies and medi-
cation underuse with physical and psychosocial health status
among a cohort of participants with arthritis.
Participants and Methods
The current study, Individual and Community Social Deter-
minants of Arthritis Outcomes Study (Social Determinants),
stems from the North Carolina Family Medicine Research
Network cohort . Briefly, in 2004 and 2005, of the 4,442
cohort members assessed for eligibility, 4,165 members
were deemed eligible (those ≥18 years of age, who agreed
to be contacted for future studies, had a current address and
telephone number, and spoke English fluently) and invited
by mailed letter to participate in the Social Determinants
study. Of the 2,479 individuals who consented and partici-
pated in the baseline telephone survey (Social Determinants
T1), 2,420 individuals agreed to be contacted again and
were subsequently mailed invitations to participate in a
follow-up telephone survey (Social Determinants T2). In
2006, a total of 1,541 participants consented to participation
and were queried about demographics, chronic conditions,
health attitudes and beliefs, and prescription medication
costs (Fig. 1); the telephone survey lasted approximately
45 min. All study materials and methods were approved
by the University of North Carolina at Chapel Hill Biomed-
ical Institutional Review Board.
This paper cross-sectionally examines the 729 partici-
pants self-reporting arthritis, as well as self-reporting cur-
rently taking doctor-prescribed medicines for their health in
the follow-up Social Determinants T2 survey. Arthritis sta-
tus was established according to the 2002 arthritis module of
the Behavioral Risk Factor Surveillance System (BRFSS)
and includes any type of doctor-diagnosed arthritis, such as
ann. behav. med.
osteoarthritis, rheumatoid arthritis, gout, lupus, or fibromy-
algia . Of the 729 individuals self-reporting arthritis, the
majority (54 %) report two or more forms of arthritis, with
the most common being osteoarthritis (n0390), bursitis/
tendonitis (n0345), rheumatoid arthritis (n0168), carpel
tunnel syndrome (n0160), gout (n0109), fibromyalgia
(n090), and other forms (n090). These respondents were
similar in race (77 % white vs. 76 %), gender (75 % female
vs. 71 %), and education (55 % some college or more vs.
50 %) when compared to the initial cohort in 2001, though
were older (25 % aged 65+ vs. 18 %). When compared to
two American samples of individuals with arthritis that were
drawn from nationally representative datasets [34,35], the
current sample had a greater percentage of non-Hispanic
black participants, and was more likely to be female, older,
more highly educated, have greater body mass index (BMI)
and greater number of comorbid conditions (data not
shown). However, these variations may reflect the differ-
ence of sampling frames, as individuals who visit family
practices in North Carolina may be different (i.e., sicker)
than individuals participating in nationally representative
Health Assessment Questionnaire
The Health Assessment Questionnaire (HAQ) measures
self-reported disability in daily function by assessing 20
activities of daily living organized around eight domains:
dressing and grooming, arising, eating, walking, hygiene,
reach, grip, and outside activities. Level of difficulty for
each is assessed on a scale from 0 (no difficulty) to 3 (unable
to do). Domain scores were summed (range 0–24) and
divided by 8 to provide a continuous, averaged index value
from 0 to 3 . In this study, the HAQ had high internal
consistency (Cronbach’sα00.94). A higher HAQ score
indicates greater disability, and previous research has
Fig. 1 Participant recruitment
ann. behav. med.
demonstrated average scores of 0.49 in a general
population-based study, 0.80 among osteoarthritis patients,
and 1.2 in rheumatoid arthritis patients [37–39].
Physical and Mental Health Functioning
The Medical Outcomes Study’s 12-item Short Form Survey
Instrument (SF-12v2) two summary scores, the SF-12v2
Physical Component (PCS), and the SF-12v2 Mental Com-
ponent (MCS) were used to assess physical and mental
health functioning. The SF-12v2 is strongly correlated with
the SF-36 and is reliable in general populations . In this
study, it had high internal consistency (Cronbach’sα00.90).
PCS and MCS summary scores were designed to range from
0 to 100, with a general US population mean score equal to
50 (standard deviation of 10) [41,42]. PCS score of 38.3
and MCS score of 48.9 have previously been noted in a
sample of southern US community-dwelling adults with
self-report arthritis . Generally, higher PCS and MCS
summary scores indicate better health.
The Centers for Disease Control (CDC) Health-Related
Quality of Life measure of global self-rated health was used
to examine participant self-report of overall health . One
question asked “in general, would you say that your health
is excellent, very good, good, fair, or poor”. Response
options were collapsed into two categories for analyses in
this study: excellent/very good/good (referent) and fair/poor.
The CDC health-related quality of life measure has shown
good construct validity, concurrent validity, and predictive
validity, and has been validated against other objective
health-related quality of life instruments [45,46].
The Center for Epidemiologic Studies Depression (CES-D)
Scale measures symptoms associated with depression in the
general population  and is a 20-item, self-report scale
yielding scores ranging from 0 to 60, with higher scores
indicating greater levels of depressive symptoms. Scores of
16 or greater have been found to be a marker of depressive
symptoms in the general population [47,48], therefore
depressive symptom scores were dichotomized at cut point
of 16 [<16 (referent) and ≥16] in this study. This scale had
high internal consistency in this study (Cronbach’sα00.92).
The Helplessness Subscale of the Rheumatology Attitudes
Index (RAI) assessed personal beliefs and attitudes regard-
ing a self-identified condition (physical or mental) that
limited activities. This is a short, five-item group of state-
ments asked in the Rheumatology Attitudes Index  that
were originally adapted from the Arthritis Helplessness
Index [50,51]. Individuals respond to each statement using
a five-point Likert scale that ranges from strongly disagree
 to strongly agree . Responses are summed and aver-
aged to create a mean score that ranges from 1 to 5, with
lower scores indicating greater perceived helplessness. This
scale had moderate internal consistency in this study (Cron-
bach’sα00.74), which is consistent with previous research
conducted in the development of this brief measure .
We asked participants to respond “yes”or “no”to whether
or not during the last 12 months they had “spent less on
food, heat or other basic needs so that you would have
enough money for your medicines,”“ever have to borrow
money from a friend or relative outside of your household to
pay for your prescription medications,”“increased the
amount of credit card debt you carried month-to month
because of the cost of your prescription medications,”or
“taken fewer medications than prescribed by your doctor
because of the cost.”These questions have been used pre-
viously in studies examining patient strategies to cope with
high-prescription medication costs, including restricting
medication use [9,21,24,53,54]. Previous research has
considered these four questions to represent three strategies:
“cutting back on necessities,”“increasing debt,”and “med-
ication restriction,”with “increasing debt”representing a
combined response of borrowing money from a friend or
relative and increasing credit card debt. For this study, we
considered these four questions to be representing four
separate strategies, given the personal nature of humbling
oneself to asking for monetary help from family or friends,
vs. independently using available credit to purchase
In this study, covariates and potential confounders included
participant sociodemographics (age, race, and gender) as
well as health characteristics (BMI and number of comorbid
conditions). Age was calculated from the date of telephone
survey and self-reported date of birth, and used as a contin-
uous measure. Race was self-reported and based on the
2000 US Census race and ethnicity categories and tricho-
tomized into non-Hispanic White (referent), non-Hispanic
Black, and other, where all else were labeled other. Educa-
tional attainment was assessed with seven categories and
later dichotomized as High School (HS)/General Education-
al Development test or below and some college and above
(referent). Household income was assessed by asking
ann. behav. med.
participants the following question: “Is your annual family
income above or below US$45,000.”This dichotomy is
retained for this study, with greater than US$45,000 per year
as the referent. Participants provided a description of their
current occupation (or last occupation if not currently
employed at time of survey) and coded according to occu-
pation classification categories from the 2000 US Census.
Occupation was further refined into two categories for this
study: nonprofessional (e.g., farming, fishing, service, con-
struction, production, and labor) and professional (e.g.,
management, technical, sales and office; referent). Home-
ownership was assessed by asking participants: “Do you
own your home?”(yes, no) with homeowner as the referent.
BMI (in kilograms per square meters) was calculated from
self-reported height and weight, and used as a continuous
measure. Existing comorbid conditions were assessed by
asking participants if a health professional ever told them
they had any of 21 different chronic diseases (e.g., asthma,
diabetes, high cholesterol). For this paper, the number of
comorbid conditions is a sum of all self-reported conditions.
After excluding for missing cases, all statistical analyses
were conducted on 729 individuals with self-report arthritis.
To examine our first aim, we conducted univariate analyses
to generate descriptive statistics, as well as correlation and
unadjusted bivariate analyses to examine the relationship
between demographics, medication cost-coping strategies,
and health status. Chi-square analyses were conducted with
categorical variables, and ttest and ANOVA were conducted
for continuous variables. We estimated multivariate logistic
regression models to identify the independent effect of char-
acteristics on using prescription medication cost-coping
strategies. To examine our second aim, we conducted sepa-
rate multivariate linear regressions (SF12v2 physical and
mental functioning, HAQ disability, and helplessness) and
multivariate logistic regressions (self-rated health and CES-
D depressive symptoms) to examine the association be-
tween prescription medication cost-coping strategies and
health status, models adjusted for age, gender, BMI, comor-
bid condition count, race, educational attainment, income,
homeownership, and occupation. Though the data are cross-
sectional, we also wanted to further examine the complex
relationship between key demographic variables, medica-
tion cost-coping strategies, and health status. We examined
whether medication cost-coping strategies mediated the as-
sociation between health status and (1) income, (2) race, and
(3) comorbid condition count by following established
criteria [55,56]. We used adjusted multivariate models
to first establish existence of a significant direct effect
(either linear or logistic depending on the health status
variable). We used logistic regression models to estab-
lish significant associations between the mediators
(medication cost-coping strategies) and the three key
demographic variables, as well as linear/logistic regres-
sion models to establish significant associations between
the mediators and health status. In each of the models,
the mediators were treated as binary variables. Further
tests of mediation (e.g., Sobel test) were not performed
based on results of the mediation analyses. All analyses
were conducted using the STATA 11.0 (StataCorp, Col-
lege Station, TX).
The 729 participants with arthritis that were currently taking
doctor-prescribed medicines for their health were on aver-
age 61 years old, had a mean BMI of 30, and had 6
comorbid conditions. They tended to be female (75 %),
non-Hispanic white (77 %), with an income below US
$45,000 (56 %), educated (55 % some college or higher),
and worked in occupations considered “professional”
(59 %). Please see Table 1. Participants had mean scores
of 37.9 and 51.5 for physical and mental health functioning,
respectively, and generally reported modest disability (mean
score 0.70). The majority of participants self-reported de-
pressive symptoms scores of less than 16 (69 %) and good
self-rated health (61 %).
When asked whether or not they had used a medication
cost-coping strategy within the last 12 months, 22 % of
participants reported spending less on basic necessities,
16 % reported borrowing money from family or friends,
12 % reported increasing credit card debt, and 20 % reported
taking fewer medications than prescribed by their doctor
(Table 1). While 65 % of participants reported not using
any strategy, nearly 35 % reported pursuing at least one
strategy (n0255): most participants reporting only one cop-
ing strategy (n0105), however some reported two (n066)
and three strategies (n067), and several reported using all
four strategies (n019).
We conducted unadjusted bivariate analyses (not shown)
and found that, in general, the following characteristics were
significantly (p<0.05) more likely to be associated with
engaging in medication cost-coping strategies: being fe-
male, younger age, higher BMI, having more comorbid
conditions, non-Hispanic Black race, nonprofessional occu-
pation, household incomes of less than US$45,000, and
being a homeowner. The exceptions were associations of
spending less on basic necessities with age (p00.062), as
well as restricting medication use with BMI (p00.058), race
(p00.691), and occupation (p00.089).
ann. behav. med.
When we examined participant sociodemographic char-
acteristics associated with medication cost-cutting strategies
through multivariate logistic regression models controlling
for covariates (Table 2), we found that individuals who were
younger were at greater odds than older adults to engage in
all cost-coping strategies. Individuals with a greater comor-
bid condition count were at greater odds of engaging in
cutting necessities (OR 1.26, 95 % CI 1.17–1.36), borrow-
ing money (OR 1.25, 95 % CI 1.15–1.36), increasing credit
card debt (OR 1.27, 95 % CI 1.16–1.39), and restricting
medication use (OR 1.27, 95 % CI 1.18–1.38). Participants
earning less than US$45,000 had about five times greater
odds than those with higher earnings to cut necessities (OR
4.76, 95 % CI 2.77–8.20) and borrow money (OR 5.14,
95 % CI 2.62–10.10), and they had nearly three times
greater odds of restricting medication use (OR 2.95, 95 %
CI 1.77–4.91). Non-Hispanic Blacks had two times greater
odds of reporting cutting necessities than non-Hispanic
Whites (OR 2.24, 95 % CI 1.37–3.64) and nearly three
times greater odds of borrowing money (OR 2.87, 95 %
CI 1.67–4.93). Of note, individuals that were homeowners
had significantly greater odds than non-homeowners of
reporting increasing credit card debt. Finally, individuals
with a nonprofessional status occupation had two times
greater odds of reporting borrowing money from family or
friends (OR 2.05, 95 % CI 1.20–3.50).
We found through the unadjusted bivariate relationship
between each medication cost-coping strategy and each
health status outcome (data not shown) that participants
engaging in each strategy were significantly at greater odds
to have worse physical health (e.g., worse physical func-
tioning, greater HAQ disability, and worse self-rated health),
as well as worse psychological health (greater helplessness,
worse mental health, and more depressive symptoms), all at
Given the demonstrated relationship between medication
cost-coping strategies and health status, we examined the
association between all four medication cost-coping strategies
and each health status (Table 3). Individuals reporting increas-
ing credit card debt scored nearly five points lower on PCS
physical health even after adjusting for covariates (B0−4.61,
p≤0.001). Those reporting borrowing money from family or
friends, as well as reporting medication underuse scored more
than a tenth of a point higher on HAQ disability (B00.13, p0
0.049 and B00.13, p00.031, respectively). Increased credit
card debt as a medication cost-coping strategy was associated
with 2.35 times greater odds of reporting worse self-rated
health (95 % CI 1.29–4.27; p00.005) and also associated with
scoring more than a half-point higher on the RAI helplessness
scale (B00.47, p≤0.001). Participants who borrowed money
from family or friends scored nearly four points lower on
mental health status (B0−3.82, p00.003) and scored more
than four points lower on mental health status when reporting
medication underuse (B0−4.21, p≤0.001). Finally, individu-
als reporting medication underuse had 2.03 times greater odds
Table 1 Participant sociodemographic characteristics and outcomes
Age, mean±SD (years); range 60.6± 12.5; 23.5–94.6
Body mass index (kg/m
), mean±SD; range 30.2± 7.1; 15.6–64.6
Comorbid condition count, mean±SD; range 6± 3; 0–16
Non-Hispanic White 77
Non-Hispanic Black 17
<US$45,000 income 56
Some college or higher 55
Physical functioning, mean±SD;
SF12v2 PCS (0–100) 37.9± 12.8; 5.7–65.6
Mental health, mean±SD; range
SF12v2 MCS (0–100) 51.5± 11.1; 9.0–75.2
Disability, mean±SD; range
HAQ (0–3) 0.70± 0.64; 0–3
Helplessness, mean±SD; range
RAI (1–5) 3.03±0.89; 1–5
CES-D score <16 69
CES-D score ≥16 31
Good/very good/excellent 61
Prior 12-month medication cost-coping strategies
Spend less on food, heat or other basic needs so as
to have enough money for medicines
Borrow money from a friend or relative outside of
your household to pay for prescription
Increase the amount of credit card debt carried
month-to-month because of prescription medi-
Prior 12-month cost-related medication underuse
Taken fewer medications than prescribed by
doctor because of the cost
Values are the percentage unless otherwise indicated
CES-D center for epidemiologic studies depression, HAQ health as-
sessment questionnaire, RAI rheumatology attitudes index, SD standard
deviation, SF12v2 MCS short form survey instrument physical com-
ponent mental component, SF12v2 PCS short form survey instrument
ann. behav. med.
Table 2 Participant sociodemographic characteristics associated with medication cost-coping strategies, N0729
Cut necessities Borrowed money Increased debt Restricted medication use
OR 95 % CI pvalue OR 95 % CI pvalue OR 95 % CI pvalue OR 95 % CI pvalue
Age 0.96 0.95–0.98 ≤0.001 0.94 0.92–0.96 ≤0.001 0.96 0.94–0.98 ≤0.001 0.94 0.92–0.95 ≤0.001
BMI 0.99 0.96–1.02 0.515 1.00 0.83–2.98 0.162 0.99 0.96–1.03 0.620 0.98 0.96–1.01 0.254
Comorbid condition count 1.26 1.17–1.36 ≤0.001 1.25 1.15–1.36 ≤0.001 1.27 1.16–1.39 ≤0.001 1.27 1.18–1.38 ≤0.001
<US$45,000 4.76 2.77–8.20 ≤0.001 5.14 2.62–10.10 ≤0.001 1.36 0.76–2.46 0.302 2.95 1.77–4.91 ≤0.001
Non-Hispanic White Ref
Non-Hispanic Black 2.24 1.37–3.64 ≤0.001 2.87 1.67–4.93 ≤0.001 1.24 0.65–2.37 0.510 0.86 0.50–1.50 0.604
Other 1.37 0.64–2.95 0.419 1.97 0.86–4.52 0.110 0.90 0.35–2.31 0.819 0.74 0.33–1.69 0.478
Female 1.17 0.70–1.97 0.542 1.57 0.83–2.98 0.162 1.37 0.72–2.60 0.336 1.39 0.82–2.37 0.225
Some college or more Ref
HS education or less 1.12 0.72–1.76 0.605 0.81 0.48–1.37 0.432 1.06 0.62–1.82 0.822 0.84 0.54–1.32 0.453
No 1.06 0.66–1.69 0.822 1.10 0.64–1.87 0.718 0.39 0.19–0.81 0.011 0.89 0.54–1.47 0.642
Nonprofessional 1.07 0.68–1.68 0.783 2.05 1.20–3.50 0.009 1.10 0.63–1.94 0.733 1.00 0.63–1.60 0.996
Each logistic model adjusted for all variables listed in table
BMI body mass index
ann. behav. med.
of reporting depressive symptoms than those who did not
report medication underuse (95 % CI 1.25–3.30; p00.004).
In general, being older and having more comorbid conditions,
higher BMI, and having lower educational attainment and
income levels were statistically significant covariates in the
majority of these models.
Finally, we conducted mediation analyses to further test
the complex relationship between each medication cost-
coping strategy and health status, income, race, and comor-
bid condition count. First, while we established statistically
significant direct effects in adjusted models for income and
health status, as well as comorbid condition count and health
status, none existed for race. Therefore, mediation analyses
were not carried out for race. The tests of mediation, based
on established criteria, revealed that none of the medication
cost-coping strategies singularly acted as mediators between
income and health status or between comorbid condition
count and health status. In all the tests of mediation, the
indirect effect remained significant and the indirect effect
beta coefficient values were comparable to the direct effect
models. For example, while low-income individuals were
more likely to engage in spending less on necessities, and
those individuals spending less on necessities had lower
physical functioning score, spending less on necessities
did not more fully explain the relationship between income
and physical functioning.
Our study is one of the first to examine medication cost-
coping strategy use and cost-related medication nonadher-
ence among individuals with self-report arthritis. We found
that, within this community-based sample, there is a high
rate of individuals with arthritis employing strategies to
cope with prescription medication cost, and 20 % of the
total sample reported restricting medication use. These rates
are on par with other rates previously identified in studies
among individuals with chronic conditions (e.g., diabetes),
low income, and older adults [20,21,57–59].
Our findings reveal that, with the exception of age and
homeownership, those typically considered at greater risk
for poorer health had engaged in cost-related coping strate-
gies in the past 12 months. Non-Hispanic Blacks and those
with a greater number of comorbid conditions, low house-
hold income, and nonprofessional occupations routinely
engaged in either cutting back on basic necessities, borrow-
ing money from family or friends, increasing credit card
debt, and/or restricting medication use. Older adults were
less likely than their younger counterparts to engage in these
strategies; however, this relationship has been previously
observed and documented in the literature [20,21]. This
study supports previous research suggesting that women
were more likely than men to cut back on necessities and
restrict medication use [20,21]. Finally, of interest is the
finding that homeowners were more apt to increase credit
card debt as a response to medication costs. This covariate
has not been previously examined in the literature, but
reveals that homeowners may be more vulnerable to
medication cost issues than non-homeowners. We hypothe-
size that these participants might either have additional
costs associated with homeownership or not qualify for
public or private medication assistance programs be-
cause of their homeownership status, and resort to using
available credit to manage their medical costs. Also for
consideration is the possibility that homeowner status
may be a proxy for overall available wealth and finan-
cial resources. This status may also have a positive
bearing on individual credit or credit scores, leaving
homeowners with the option of choosing to increase
credit card debt more often than non-homeowners.
It is important to consider our results in the context of
what might be considered a minimally important difference
in health status, such as physical functioning or disability.
Prior research suggests that a universal value of effect size
equal to 0.5 is the minimally important difference that a
Table 3 Strategies to cope with medication costs, socioeconomic status variables, and health status, beta (SE), and OR [95 % CI], N0729
Physical health Psychological health
Physical functioning Disability Self-rated health Helplessness
Mental health Depressive symptoms
B(SE) B(SE) OR (95 % CI) B(SE) B(SE) OR (95 % CI)
Spent less on necessities −1.54 (1.23) 0.08 (0.06) 0.99 (0.59, 1.62) 0.11 (0.12) 0.47 (1.15) 1.39 (0.86, 2.26)
Borrowed money 1.01 (1.37) 0.13 (0.07)* 0.93 (0.53, 1.65) 0.08 (0.13) −3.82 (1.29)** 0.95 (0.55, 1.64)
Increased credit card debt −4.61 (1.41)*** 0.06 (0.07) 2.35 (1.29, 4.27)** 0.47 (0.13)*** −0.01 (1.33) 0.81 (0.45, 1.45)
Medication underuse 0.03 (1.21) 0.13 (0.06)* 1.61 (0.97, 2.66) −0.11 (0.11) −4.21 (1.14)*** 2.03 (1.25, 3.30)**
Models adjust for age, gender, body mass index, comorbid condition count, race, education, income, homeownership, and occupation
*p<0.05; **p< 0.01; ***p≤0.001
ann. behav. med.
patient would report , however this value is likely to
differ given the instruments and various scale anchors used
. For example, a previously reported physical function
(SF-36 PCS) minimally important difference corresponds to
an effect size of 0.49 , whereas a change in disability
(HAQ) of about 0.20 has been reported as clinically impor-
tant [39,63], corresponding to an effect size of 0.27. Expert
opinion varies widely on what constitutes a meaningful
difference and its assessment. Our group has previously
calculated quasi-effect size for health status  to better
interpret results. We similarly calculated quasi-effect size for
physical functioning, disability, helplessness, and mental
health by standardizing the parameter estimates by
corresponding standard deviations in this study. We interpret
the parameter estimates as changes in health status resulting
from a treatment of medication cost-coping strategy use
relative to the referent group (strategy not used). For exam-
ple, the PCS parameter estimate for increasing credit card
debt is −4.61, meaning that people employing this strategy
would have an average PCS that is 4.61 units less than those
not employing the strategy. Using the sample standard de-
viation of PCS (12.8) from Table 1, a quasi-effect size
would be 4.61/12.800.36. The quasi-effect size for disabil-
ity (HAQ) is 0.20 for strategies “borrowing money”and
“medication underuse,”and a quasi-effect size of 0.53 for
helplessness among those who increased credit card debt.
Finally, the quasi-effect size was 0.34 and 0.38 for mental
health status for those employing the strategies “borrowing
money”and “medication underuse.”Our findings represent
moderate effects judged against other estimates of minimal-
ly important difference, suggesting that use of various med-
ication cost-coping strategies, including medication
underuse, may result in poorer health status for individuals
Medication underuse or nonadherence of nonsteroidal
anti-inflammatory drugs (NSAIDs), analgesics, disease-
modifying antirheumatic drugs (DMARDs), and biologics
can result in ineffective management of symptoms (e.g.,
pain, fatigue), inflammation and swelling, progression of
joint damage, and hospitalization [19,65–67]. It is possible
that without adherence to a proper medical regime, mobility
and function have been compromised—influencing both
physical functioning and eventually quality of life in our
study population. Failure to comply with dosage on pre-
scription medications for comorbid conditions (e.g., heart
disease, depression) may have contributed to poorer physi-
cal and psychosocial health status among those with
Medication cost-coping strategies that involved cutting
back on necessities or needing additional financial resources
to afford prescriptions were also associated with worse
physical and psychosocial health. It is possible that engag-
ing in these behaviors (e.g., purchasing fewer groceries,
skipping a bill payment 1 month or asking for financial
help) may increase the level of stress and anxiety stemming
from not having sufficient resources for even basic neces-
sities or health needs, and in turn negatively influence
individual health. It should be noted, however, that earlier
research has demonstrated that individuals who are sicker or
take more medications often engage in medication cost-
coping strategies . While our analyses in this study did
not reveal medication cost-coping strategies acting as medi-
ators in the relationship between income and health status or
in the relationship between comorbid condition count and
health status, this type of potentially reciprocal relationship
deserves additional consideration. This study is cross-
sectional and cannot establish a causal link between cost-
coping strategies and health status. Future studies might
consider examining the longitudinal impact of medication
cost-coping strategies on disease burden and general health
among individuals with different forms of arthritis because
of the varying disease-associated prescription medication
costs (e.g., lower cost of NSAIDs for osteoarthritis vs.
higher cost of DMARDs for rheumatoid arthritis). If certain
groups engage in medication cost-coping strategies over a
prolonged period of time, medication cost-coping strategies
and nonadherence may contribute to the relationship be-
tween characteristics that put individuals at greater risk for
poorer health (e.g., low income) and poorer health outcomes
(e.g., physical and mental health functioning). Future re-
search might also focus on the role of stress and anxiety as
potential mediators in the relationship between engaging in
medication cost-coping strategies and health outcomes.
Our study has several limitations that should be noted in
light of our findings. First, there is potential for reporting
bias to act in complex ways in this study. Participants may
have been generally reluctant to report that they have diffi-
culties affording their prescription medications. For exam-
ple, individuals reluctant to report using cost-coping
strategies may have caused the strength of the associations
to be reduced, suggesting that the use of cost-coping strat-
egies is more widespread than reported. On the other hand,
reporting bias may have also varied according to demo-
graphic groups such as race, gender, and health status. For
example, if a particular demographic felt more at liberty to
report using cost-coping strategies or a particular cost-
coping strategy than another, the observed relationships
may have been greater due to reporting bias. The unquanti-
fied level, direction, and complexity of reporting bias should
be considered when interpreting these data.
Second, this study does not have information on other
important variables, such as health insurance status or out-
of-pocket health care costs. While nearly 40 % of partici-
pants were 65 or older, and therefore eligible for Medicare,
we did not infer insurance status for this subsection of the
sample. However, it is possible that Medicare eligibility
ann. behav. med.
may offer a partial explanation for the observed age differ-
ences in strategy use: that is to say older individuals did not
engage in medication cost-coping strategies because they
had some form of health insurance. Individuals with greater
out-of pocket health care costs may be more likely to engage
in cost-coping strategies than those with lower out-of pocket
costs. Unfortunately we are not able to examine the role of
out-of pocket prescription medication costs, either through
adjustment or sample stratification, and we must reiterate
that our findings are largely descriptive given these limita-
tions. However, studies have shown that individuals who are
uninsured or underinsured have greater cost-related poor
adherence [9,69], and it could be through this mechanism
that we observed the association between cost-related med-
ication nonadherence and poorer physical and psychosocial
Additionally, this study did not ask participants how
many medications they were taking, about medication dos-
ing complexity, or whether the medications individuals were
using were disease specific (e.g., heart disease, arthritis) or
for general health (e.g., multivitamin) or if they were using
complementary and alternative medicine to manage their
arthritis symptoms (e.g., liniments, fish oil). We also did
not ask participants about their perceived beliefs regarding
the importance of arthritis-related medication adherence vs.
medication adherence for other chronic conditions. While
previous research found no increase in disease-specific com-
plications among individuals with arthritis , it has estab-
lished that individuals do cut back on symptom-relieving
medications (e.g., nonsteroidal anti-inflammatory drugs)
rather than medications for life-threatening conditions such
as diabetes or heart disease [10,70]. Future research might
qualitatively examine which medications arthritis patients
specifically cut back on when costs prohibit medication
adherence, as well as further investigate beliefs and attitudes
towards medication efficacy. Finally, although prior re-
search has suggested the importance of trusting patient–
provider interactions regarding medication costs and ways
to limit medication underuse due to cost , we did not
have data to explore these additional associations among
individuals with arthritis.
This study supports the existing literature and also provides
a foundation for future studies to examine the complex rela-
tionship between financial burden, complexity of medication
adherence, individual characteristics (e.g., race, age, health
conditions), interactions with health care providers around
medication adherence and cost, and the health care system
among individuals with arthritis [13,72]. It should be high-
lighted again that this study asked North Carolinians about
medication cost-coping strategies in 2006, and therefore may
not be generalizable to other populations. However, given the
current global economic climate, these issues may be of
particularrelevance in other geographic areas, and use of these
strategies may be even more extreme and widespread then
reported in our 2006 study. Our findings also reinforce that
primary care providers, as well as rheumatologists, need to be
attuned to the costs of medication they prescribe and consider
identifying whether medication costs (e.g., out-of-pocket
costs) present a financial burden or additional stress to their
patients [53,73]. Knowledge of how individuals with arthritis
choose to cope with medication costs can present opportuni-
ties for interventions ranging from networking patients with
pharmaceutical benefits programs, to conducting an audit of
all current prescriptions, to reinforcement of the importance of
medication adherence. In addition, physical activity program-
ming and health promotion campaigns (e.g., the Arthritis
Foundation’s“Moving is the Best Medicine”) are aimed at
decreasing incidence rates, as well as reducing symptoms and
disability resulting from prevalent arthritis. These types of
interventions can act as tertiary prevention and may reduce
the number of individuals who experience poor overall health
and/or arthritis-related outcomes due to engaging in medica-
tion cost coping.
In conclusion, this descriptive study demonstrates that
medication cost-coping strategies are independently and
significantly associated with both physical and psychosocial
health status among individuals with arthritis. Individuals
who engage in medication cost-coping strategies did in fact
have poorer health, independent of known risk factors for
poor health in this population (e.g., age, gender, comorbid
conditions, low income). Researchers, practitioners, public
health practitioners, and policy makers should strive for
continued collaboration in efforts to reduce the overall bur-
den of chronic disease and arthritis, particularly those relat-
ed to high costs of prescription medications.
Acknowledgments The authors thank the following participating
family practices in the NC-FM-RN for their assistance: Black River
Health Services, Burgaw; Bladen Medical Associates, Elizabethtown;
Blair Family Medicine, Wallace; Cabarrus Family Medicine, Concord;
Cabarrus Family Medicine, Harrisburg; Cabarrus Family Medicine,
Kannapolis; Cabarrus Family Medicine, Mt. Pleasant; Chatham Prima-
ry Care, Siler City; Carolinas Medical Center Biddle Point, Charlotte;
Carolinas Medical Center North Park, Charlotte; Community Family
Practice, Asheville; Cornerstone Medical Center, Burlington; Day-
spring Family Medicine, Eden; Family Practice of Summerfield, Sum-
merfield; Goldsboro Family Physicians, Goldsboro; Henderson Family
Health Center, Henderson; Orange Family Medical Group, Hillsbor-
ough; Person Family Medical Center, Roxboro; Pittsboro Family Med-
icine, Pittsboro; Prospect Hill Community Health Center, Prospect
Hill; Robbins Family Practice, Robbins; and Village Family Medicine,
Chapel Hill. Finally, we thank the individuals who willingly partici-
pated in the study.
Grant Support This project was supported by a grant from the
National Institute of Arthritis and Musculoskeletal and Skin Disease
Multidisciplinary Clinical Research Center Rheumatic Diseases: P60-
AR049465-05 and the Thurston Arthritis Research Center Training
Grant 5T32-AR007416. This research was conducted while Dr. Martin
was a postdoctoral research fellow at the Thurston Arthritis Research
Center. She is currently a National Institute on Aging (NIA) postdoctoral
ann. behav. med.
research fellow, and her research is now supported in part by the Intra-
mural Research Program of the NIA.
Conflict of Interest Statement The authors have no conflict of
interest to disclose.
1. Centers for Disease Control and Prevention (CDC). Prevalence of
doctor-diagnosed arthritis and arthritis-attributable activity limita-
tion—United States, 2003–2005. MMWR. 2006; 55(40): 1089–
2. Abraido-Lanza AF, Revenson TA. Illness intrusion and psycho-
logical adjustment to rheumatic diseases: A social identity frame-
work. Arthritis Rheum. 2006; 55(2): 224–232.
3. Katz PP, Morris A. Use of accommodations for valued life activ-
ities: Prevalence and effects on disability scores. Arthritis Care
Res. 2007; 57(5): 730–737.
4. Centers for Disease Control and Prevention (CDC). Prevalence
and most common causes of disability among adults—United
States, 2005. MMWR. 2009; 58(16): 421–426.
5. Centers for Disease Control and Prevention (CDC). National and
state medical expenditures and lost earnings attributable to arthritis
and other rheumatic conditions—United States, 2003. MMWR.
2007; 56(1): 4–7.
6. Yelin E, Murphy L, Cisternas MG, Foreman AJ, Pasta DJ, Hel-
mick CG. Medical care expenditures and earnings losses among
persons with arthritis and other rheumatic conditions in 2003, and
comparisons with 1997. Arthritis Rheum. 2007; 56(5): 1397–1407.
7. Cisternas MG, Murphy LB, Yelin EH, Forman AJ, Pasta DJ,
Helmick CG. Trends in medical care expenditures of US adults
with arthritis and other rheumatic conditions 1997 to 2005. J
Rheumatol. 2009; 36(11): 2531–2538.
8. Dunlop DD, Manheim LM, Song J, Chang RW. Health care
utilization among older adults with arthritis. Arthritis Rheum.
2003; 49(2): 164–171.
9. Mojtabai R, Olfson M. Medication costs, adherence, and health
outcomes among Medicare beneficiaries. Health Aff. 2003; 22(4):
10. Piette JD, Heisler M, Wagner TH. Cost-related medication under-
use among chronically III adults: The treatments people forgo, how
often, and who is at risk. Am J Public Health. 2004; 94(10): 1782–
11. Kaufman DW, Kelly JP, Rosenberg L, Anderson TE, Mitchell AA.
Recent patterns of medication use in the ambulatory adult popula-
tion of the United States. JAMA. 2002; 287(3): 337–344.
12. Stagnitti M. Trends in outpatient prescription drug utilization and
expenditures, 1997 and 2004. Agency for Healthcare Research and
Quality [cited 2010 Sept. 29]; April 2007. Available at http://
13. Briesacher BA, Gurwitz JH, Soumerai SB. Patients at-risk for cost-
related medication nonadherence: A review of the literature. J Gen
Intern Med. 2007; 22(6): 864–871.
14. National Center for Health Statistics. Health, United States, 2009
with Special Feature on Medical Technology. Hyattsville, Mary-
land: Public Health Service; 2009.
15. Cohen E. Ten ways to save on prescription drugs. CNNhealth com
[cited 2010 May 20]. Available at http://articles.cnn.com/2009-03-
16. Sticker shock at the pharmacy counter: CR poll finds high cost of
drugs leads to risky behaviors. Consumer Reports Health [cited
2010 May 20]. Available at http://www.consumerreports.org/
17. Connolly C, Marr K. As budgets tighten, more people decide
medical care can wait. The Washington Post 2008 Oct 16.
18. Donovan JL, Blake DR. Patient non-compliance: Deviance or
reasoned decision-making? Social Sci Med. 1992; 34(5): 507–513.
19. Elliott RA. Poor adherence to medication in adults with rheuma-
toid arthritis: Reasons and solutions. Dis Manag Health Outcomes.
2008; 16(1): 13–29.
20. Heisler M, Wagner TH, Piette JD. Patient strategies to cope with
high prescription medication costs: Who is cutting back on neces-
sities, increasing debt, or underusing medications? J Behav Med.
2005; 28(1): 43–51.
21. Piette JD, Heisler M, Wagner TH. Problems paying out-of-pocket
medication costs among older adults with diabetes. Diabetes Care.
2004; 27(2): 384–391.
22. Dunlay SM, Eveleth JM, Shah ND, McNallan SM, Roger VL.
Medication adherence among community-dwelling patients with
heart failure. Mayo Clin Proc. 2011; 86(4): 273–281.
23. Zivin K, Ratliff S, Heisler MM, Langa KM, Piette JD. Factors
influencing cost-related nonadherencetomedicationinolder
adults: A conceptually based approach. Value Health. 2010; 13
24. Heisler M, Langa KM, Eby EL, Fendrick AM, Kabeto MU, Piette
JD. The health effects of restricting prescription medication use
because of cost. Med Care. 2004; 42(7): 626–634.
25. Heisler M, Choi H, Rosen AB, Vijan S, Kabeto M, Langa KM, et
al. Hospitalizations and deaths among adults with cardiovascular
disease who underuse medications because of cost: A longitudinal
analysis. Med Care. 2010; 48(2): 87–94.
26. Piette JD, Wagner TH, Potter MB, Schillinger D. Health insurance
status, cost-related medication underuse, and outcomes among
diabetes patients in three systems of care. Med Care. 2004; 42
27. Wilson IB, Rogers WH, Chang H, Safran DG. Cost-related skip-
ping of medications and other treatments among Medicare benefi-
ciaries between 1998 and 2000. J Gen Intern Med. 2005; 20(8):
28. Tamblyn R, Laprise R, Hanley JA, Abrahamowicz M, Scott S,
Mayo N, et al. Adverse events associated with prescription drug
cost-sharing among poor and elderly persons. JAMA. 2001; 285
29. Harrold LR, Andrade SE. Medication adherence of patients with
selected rheumatic conditions: A systematic review of the litera-
ture. Semin Arthritis Rheum. 2009; 38(5): 396–402.
30. Curkendall S, Patel V, Gleeson M, Campbell RS, Zagari M,
Dubois R. Compliance with biologic therapies for rheumatoid
arthritis: Do patient out of pocket payments matter? Arthritis Care
Res. 2008; 59(10): 1519–1526.
31. Garcia Popa-Lisseanu MG, Greisinger A, Richardson M, O’Malley
KJ, Janssen NM, et al. Determinants of treatment adherence in
ethnically diverse, economically disadvantaged patients with rheu-
matic disease. J Rheumatol. 2005; 32(5): 913–919.
32. Sloane PD, Callahan L, Kahwati L, Mitchell CM. Development of
a practice-based patient cohort for primary care research. Fam
Med. 2006; 38(1): 50–57.
33. Centers for Disease Control and Prevention (CDC). BRFSS Ar-
thritis Questions 1996–2010. Centers for Disease Control and
Prevention [cited 2010 July 27]. Available at http://www.cdc.gov/
34. Shih M, Hootman JM, Kruger J, Helmick CG. Physical activity in
men and women with arthritis: National Health Interview Survey,
2002. Am J Prev Med. 2006; 30(5): 385–393.
ann. behav. med.
35. Dunlop DD, Semanik P, Song J, Manheim LM, Shih V, Chang
RW. Risk factors for functional decline in older adults with arthri-
tis. Arthritis Rheum. 2005; 52(4): 1274–1282.
36. Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient
outcome in arthritis. Arthritis Rheum. 1980; 23(2): 137–145.
37. Bruce B, Fries JF. The Stanford Health Assessment Questionnaire:
Dimensions and practical applications. Health Qual Life Outcomes
38. Bruce B, Fries JF. The Stanford Health Assessment Questionnaire:
A review of its history, issues, progress, and documentation. J
Rheumatol. 2003; 30(1): 167–178.
39. Pope JE, Khanna D, Norrie D, Ouimet JM. The minimally impor-
tant difference for the Health Assessment Questionnaire in rheu-
matoid arthritis clinical practice is smaller than in randomized
controlled trials. J Rheumatol. 2009; 36(2): 254–259.
40. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health
Survey: Construction of scales and preliminary tests of reliability
and validity. Med Care. 1996; 34: 220–233.
41. Ware JE Jr, Kosinski M, Keller SD. SF-36 Physical and Mental
Health Summary Scales: A User’s Manual. Boston: The Health
Institute, New England Medical Center; 1994.
42. Gandhi SK, Salmon JW, Zhao SZ, Lambert BL, Gore PR, Conrad
K. Psychometric evaluation of the 12-item short-form health sur-
vey (SF-12) in osteoarthritis and rheumatoid arthritis clinical trials.
Clin Ther. 2001; 23(7): 1080–1098.
43. Kovac SH, Mikuls TR, Mudano A, Saag KG. Health-related
quality of life among self-reported arthritis sufferers: Effects of
race/ethnicity and residence. Qual Life Res. 2006; 15(3): 451–460.
44. Centers for Disease Control and Prevention. Measuring Healthy
Days: Population Assessment of Health-Related Quality of Life.
Atlanta: CDC; 2000.
45. Moriarty DG, Kobau R, Zack MM, Zahran HS. Tracking healthy
days—A window on the health of older adults. Prevent Chron Dis.
2005; 2(3): 1–8.
46. Mielenz T, Jackson E, Currey S, DeVellis R, Callahan LF. Psy-
chometric properties of the Centers for Disease Control and Pre-
vention Health-Related Quality of Life (CDC HRQOL) items in
adults with arthritis. Health Qual Life Outcomes. 2006; 4: 66.
47. Radloff LS. The CES-D scale: A self-report depression scale for
research in the general population. Appl Psychol Meas. 1977; 1:
48. Boyd JH, Weissman MM, Thompson WD, Myers JK. Screening
for depression in a community sample. Understanding the discrep-
ancies between depression symptom and diagnostic scales. Arch
Gen Psychiatry. 1982; 39(10): 1195–1200.
49. Callahan LF, Brooks RH, Pincus T. Further analysis of learned
helplessness in rheumatoid arthritis using a “Rheumatology Atti-
tudes Index”. J Rheumatol. 1988; 15(3): 418–426.
50. Nicassio PM, Wallston KA, Callahan LF, Herbert M, Pincus T. The
measurement of helplessness in rheumatoid arthritis. The develop-
ment of the arthritis helplessness index. J Rheumatol. 1985; 12(3):
51. Stein MJ, Wallston KA, Nicassio PM. Factor structure of the
arthritis helplessness index. J Rheumatol. 1988; 15(3): 427–432.
52. DeVellis RF, Callahan LF. A brief measure of helplessness in
rheumatic disease: The helplessness subscale of the Rheumatology
Attitudes Index. J Rheumatol. 1993; 20(5): 866–869.
53. Heisler M, Wagner TH, Piette JD. Clinician identification of
chronically ill patients who have problems paying for prescription
medications. Am J Med. 2004; 116(11): 753–758.
54. Gellad WF, Haas JS, Safran DG. Race/ethnicity and nonadherence
to prescription medications among seniors: Results of a national
study. J Gen Intern Med. 2007; 22(11): 1572–1578.
55. Baron RM, Kenny DA. The moderator-mediator variable distinc-
tion in social psychological research: Conceptual, strategic, and
statistical considerations. J Pers Soc Psychol. 1986; 51: 1173–
56. MacKinnon DP. Introduction to Statistical Mediation Analyses.
New York, NY: Taylor & Francis Group; 2008.
57. Steinman MA, Sands LP, Covinsky KE. Self-restriction of medi-
cations due to cost in seniors without prescription coverage. J Gen
Intern Med. 2001; 16(12): 793–799.
58. Safran DG, Neuman P, Schoen C, Montgomery JE, Li W, Wilson
IB et al. Prescription drug coverage and seniors: How well are
states closing the gap? Health Aff 2002.
59. Kennedy J, Erb C. Prescription noncompliance due to cost among
adults with disabilities in the United States. Am J Public Health.
2002; 92(7): 1120–1124.
60. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in
health-related quality of life: The remarkable universality of half a
standard deviation. Med Care. 2003; 41(5): 582–592.
61. Hays RD, Farivar SS, Liu H. Approaches and recommendations
for estimating minimally important differences for health-related
quality of life measures. COPD: J Chron Obstr Pulm Dis. 2004; 2
62. Farivar SS, Liu H, Hays RD. Half standard deviation estimate of
the minimally important difference in HRQOL scores? Expert Rev
Pharmacoecon Outcomes Res. 2004; 4(5): 515–523.
63. Redelmeier DA, Lorig K. Assessing the clinical importance of
symptomatic improvements. An illustration in rheumatology. Arch
Intern Med. 1993; 153(11): 1337–1342.
64. Callahan LF, Martin KR,Shreffler J,Kumar D, Schoster B, Kaufman
JS, et al. Independent and combined influence of homeownership,
occupation, education, income and community poverty on physical
health in persons with arthritis. Arthritis Care Res. 2011; 63(5): 643–
65. Chambers SA, Rahman A, Isenberg DA. Treatment adherence and
clinical outcome in systemic lupus erythematosus. Rheumatology.
2007; 46(6): 895–898.
66. Pascual-Ramos V, Contreras-Yanez I, Villa AR, Cabiedes J, Rull-
Gabayet M. Medication persistence over 2 years of follow-up in a
cohort of early rheumatoid arthritis patients: Associated factors and
relationship with disease activity and with disability. Arthritis Res
Ther. 2009; 11(1): R26.
67. Rojas-Serrano J, Cardiel MH. Lupus patients in an emergency unit.
Causes of consultation, hospitalization and outcome. A cohort
study. Lupus. 2000; 9(8): 601–606.
68. Kirking DM, Lee JA, Ellis JJ, Briesacher B, Mckercher PL.
Patient-reported underuse of prescription medications: A compar-
ison of nine surveys. Med Care Res Rev. 2006; 63(4): 427–446.
69. Piette JD, Heisler M. Problems due to medication costs among VA
and non-VA patients with chronic illnesses. Am J Manag Care.
2004; 10(11(part 2)): 861–868.
70. Goldman DP, Joyce GF, Escarce JJ, Pace JE, Solomon MD, Laouri
M, et al. Pharmacy benefits and the use of drugs by the chronically
Ill. JAMA. 2004; 291(19): 2344–2350.
71. Piette JD, Heisler M, Krein S, Kerr EA. The role of patient-
physician trust in moderating medication nonadherence due to cost
pressures. Arch Intern Med. 2005; 165(15): 1749–1755.
72. Piette JD, Heisler M, Horne R, Alexander GC. A conceptually
based approach to understanding chronically ill patients’
responses to medication cost pressures. Soc Sci Med. 2006;
73. Piette JD, Heisler M, Wagner TH. Cost-related medication under-
use: Do patients with chronic illnesses tell their doctors? Arch
Intern Med. 2004; 164(16): 1749–1755.
ann. behav. med.