Research in Social and Administrative Pharmacy 14 (2018) 53e61

Contents lists available at ScienceDirect

Research in Social and Administrative Pharmacy

journal homepage: www.rsap.org

Association between community pharmacy loyalty and persistence

and implementation of antipsychotic treatment among individuals
with schizophrenia

goire a, b, c, Alain Lesage d,
Frank E. Zongo a, b, c, Jocelyne Moisan a, b, c, Jean-Pierre Gre
a, b, c a, b, c, *
Anara Richi Dossa , Sophie Lauzier
a
Chair on Adherence to Treatments, Universit
e Laval, 1050 chemin Ste-Foy, Qu
ebec, QC, Canada
b
CHU de Qu
ebec-Universit

e Laval Research Centre, Population Health and Optimal Health Practices Research Unit, 1050 chemin Ste-Foy, Qu

ebec, QC,
Canada
c
Faculty of Pharmacy, Universit
e Laval, 1050 Avenue de la M
edecine, Qu

ebec, QC, Canada
d
Centre de Recherche Fernand-Seguin, Ho ^pital Louis-H. Lafontaine, Unit

e 218, 7331 Rue Hochelaga, Montr

eal, QC, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Background: Non-adherence is a major obstacle to optimal treatment of schizophrenia. Community
Received 1 June 2016 pharmacists are in a key position to detect non-adherence and put in place interventions. Their role is
Received in revised form likely to be more efficient when individuals are loyal to a single pharmacy.
8 December 2016
Objective: To assess the association between the level of community pharmacy loyalty and persistence
Accepted 26 December 2016
with and implementation of antipsychotic drug treatment among individuals with schizophrenia.
Methods: A cohort study using databases from the Quebec health insurance board (Canada) was con-
Keywords:
ducted among new antipsychotic users insured by Quebec's public drug plan. Level of community
Schizophrenia
Antipsychotics
pharmacy loyalty was assessed as the number of pharmacies visited in the year after antipsychotics
Community pharmacy services initiation. Persistence was defined as having an antipsychotic supply in the user's possession on the 730th
Pharmaceutical services day after its initiation and implementation as having antipsychotics in the user's possession for
80% of
Medication adherence the days in the second year after antipsychotics initiation (among persistent only). Generalized linear
Continuity of patient care models were used to estimate adjusted prevalence ratios (aPR) and 95% confidence intervals (95%CI).
Results: 6,251 individuals were included in the cohort and 54.1% had their drug prescriptions filled in >1
pharmacy. When compared to those who had their prescriptions filled in a single pharmacy, those who
had their prescriptions filled in
4 different pharmacies were 22% more likely to be non-persistent
(aPR ¼ 1.22; 95%CI ¼ 1.10e1.37) and 49% more likely to have an antipsychotic for <80% of the days
(aPR ¼ 1.49; 95%IC ¼ 1.28e1.74).
Conclusion: This first exploration of community pharmacy loyalty in the context of severe mental illness
indicates that this healthcare organisation factor might be associated with antipsychotics persistence and
implementation. Identification of individuals with low community pharmacy loyalty and initiatives to
optimize community pharmacy loyalty could contribute to enhanced persistence and implementation.
© 2016 Elsevier Inc. All rights reserved.

1. Introduction

Antipsychotics represent the cornerstone of schizophrenia

List of abbreviations: RAMQ, R

egie de l'assurance-maladie du Qu
ebec; GIS, guar-
anteed income supplement; ICD, International Classification of Diseases; PDC,
management for many individuals.1 Medication adherence, that is
proportion of days covered; DDD, defined daily dose; INN, International Non- the extent to which a person's behaviour corresponds with agreed
Proprietary Names; PR, prevalence ratio; 95%CI, 95% confidence interval; SD, recommendations from a healthcare provider,2 is key to fully
standard deviation. benefit from antipsychotic treatment. Medication adherence is a
* Corresponding author. CHU de Que
bec-Universite
Laval Research Centre, Pop-
^pital du Saint-Sac-
multidimensional construct that includes initiation (the extent to
ulation Health and Optimal Health Practices Research Unit, Ho
rement, 1050 chemin Ste-Foy, Que
bec, QC, G1S 4L8, Canada. which a newly prescribed drug treatment is undertaken), persis-
E-mail address: sophie.lauzier@pha.ulaval.ca (S. Lauzier). tence (the extent to which the drug treatment is taken for the

http://dx.doi.org/10.1016/j.sapharm.2016.12.006
1551-7411/© 2016 Elsevier Inc. All rights reserved.
54 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61
prescribed duration) and implementation (the extent to which the
drug treatment is taken at the recommended dose).3,4
However, both non-persistence and suboptimal implementa-
tion are widespread among individuals prescribed antipsychotics.
In a study conducted in the province of Quebec (Canada), 32.5% of
individuals with schizophrenia insured by the public drug plan
were non-persistent with their atypical antipsychotics one year
after initiation.5 Among those who were persistent, 21.4% were
judged as having suboptimal implementation, defined as having an
antipsychotic in their possession for < 80% of the days. Based on a
review of five studies, Lacro et al.6 found that 49.5% (weighted
mean by sample) of individuals with schizophrenia did not take
their antipsychotic treatment for at least 75% of the total number of
days, for periods varying from 1 month to 2 years. In another study,
only 55% of individuals with schizophrenia were considered
persistent and reported to have taken 100% of their antipsychotics
in the week preceding the study, with antipsychotic serum con-
centration within reference values.7
Suboptimal implementation of antipsychotic treatment has been
associated with higher risk for relapse,8,9 hospitalization,10,11 and
visits to the emergency room.12,13 For example, a study showed that
individuals who had an antipsychotic for less than 80% of the total
days were 55% more likely to be hospitalized (OR 1.55; 95% CI
1.21e1.98) and 49% more likely to use emergency psychiatric ser-
vices in the following two years (OR 1.49; 95% CI 1.12e1.98) when
compared to those having an antipsychotic for
80% of the days.12
Community pharmacists are in a key position to detect non-
adherence, identify its causes, and put in place interventions.
Some trials indicate that pharmacy-based interventions could
improve adherence among individuals with mental illness.14,15 It is
likely that the pharmacist's work could be facilitated and be more
efficient when an individual is loyal to a single pharmacy meaning
that s/he refills all his/her prescriptions at the same pharmacy. In
the province of Quebec (Canada), it is mandatory for community
pharmacists to maintain a pharmacy record with detailed infor-
mation on all the prescriptions that have been dispensed to an
individual. However, these records are not necessarily shared be-
tween pharmacies or pharmacy chains. In this context, a high
community pharmacy loyalty level implies that the pharmacist has
more complete information on drugs claimed by an individual and
that s/he could more easily use these records to detect non-
adherence. In addition, a high level of community pharmacy loy-
alty could facilitate patient-pharmacist collaboration, the imple-
mentation and follow-up of the pharmacist's adherence-enhancing
interventions as well as the collaboration with other healthcare
professionals.
Fig. 1. Study design and timelines for the assessment of persistence and implementation.
Community pharmacy loyalty has been studied in the general
population16,17 and among individuals with multiple chronic ill-
nesses18,19 or a single illness such as diabetes20 or cardiovascular
disease.21 The association between community pharmacy loyalty
and implementation was assessed in four of these studies.18e21 As
opposed to individuals who had their prescriptions filled in a single
pharmacy, those who had them filled in multiple pharmacies were
more likely to have lower levels of treatment implementation.18e21
Our team has studied to what extent individuals with schizo-
phrenia had their prescriptions filled at a single pharmacy.22 Re-
sults indicated that 42.2% of individuals were dispensed their drugs
in more than one pharmacy in the year after treatment initiation.22
As an important proportion of individuals treated for schizophrenia
were filling their prescriptions in multiple pharmacies, a study was
undertaken to determine the association between community
pharmacy loyalty and persistence with and implementation of
antipsychotics. Specific objectives of this study were:
1) To estimate the association between the level of community
pharmacy loyalty and persistence with antipsychotics on the
730th day after antipsychotics initiation;
2) Among persistent individuals, to estimate the association be-
tween the level of community pharmacy loyalty and imple-
mentation of antipsychotics in the second year after
antipsychotics initiation.
2. Methods
2.1. Study design
A cohort study of individuals diagnosed with schizophrenia who
initiated antipsychotics and who were insured with the Quebec's
public drug plan administered by the Quebec health insurance
board (R
egie de l'assurance-maladie du Qu
ebec (RAMQ)) was con-
ducted. This public drug plan covers all residents who do not have a
private group plan drug insurance, individuals aged
65 years, and
welfare recipients.23
The level of community pharmacy loyalty was measured in the
365 days after antipsychotics initiation. Persistence was estimated
the 730th day after antipsychotics initiation and implementation
was assessed between the 365th day and the 730th day after anti-
psychotics initiation (Fig. 1). This design was used to ensure that
community pharmacy loyalty would be estimated over a one-year
period, and prior to the measurement of persistence and imple-
mentation. Estimating persistence at the 730th day after
 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61
antipsychotics initiation was also based on Canadian guidelines
recommending that antipsychotics should be continued minimally
for two years following first recovery of symptoms.1
2.2. Data source
The following RAMQ databases were used: the demographic
database containing socio-economic information on all benefi-
ciaries insured by the Quebec health insurance board (age, sex,
region, guaranteed-income-supplement (GIS) status), the physician
claims database containing information on ambulatory visits (date
and diagnosis code), the drug prescription claims database con-
taining information on prescription drugs claimed (drug identifi-
cation, date dispensed, number of days' supply, prescriber's
speciality, pharmacy unique number) and hospitalization database
(date and diagnosis code). The Quebec's Institute of Statistics
(Institut de la Statistique du Qu

ebec) provided date of death, when
applicable. The information for an individual available in the
different databases was linked using a unique identifier.
2.3. Subjects
Quebec public drug plan recipients who had at least one claim
for an antipsychotic (typical or atypical) between 2000/01/01 and
2005/12/31 and were eligible for the public drug plan for the entire
365 days before the date of the first antipsychotic claim were
identified by the RAMQ. The first antipsychotic claim was defined as
the date of antipsychotics initiation.
First, individuals were included if they were
20 years of age on
the date of antipsychotics initiation. Second, to make sure anti-
psychotics were used in the treatment of schizophrenia, individuals
had to have an ambulatory or hospital schizophrenia-related
diagnosis code in the 365 days before initiation or, on the date of
antipsychotics initiation based on the International Classification of
Diseases (ICD) (ICD-9: codes 295.0 to 295.9 (schizophrenic disor-
ders); ICD-10: codes F20 (schizophrenia), F21 (schizotypal disor-
der), F232 (other acute predominantly delusional psychotic
disorders) and F25 (schizoaffective disorders). These ICD codes
have been shown to have very good validity when identifying in-
dividuals with schizophrenia.24 Third, individuals had to have no
antipsychotic claim in the 365 days preceding the antipsychotics
initiation in order to include new antipsychotics users only. Fourth,
individuals had to be eligible for the Quebec public drug plan for
the entire 730 days after the date of antipsychotics initiation.
Finally, to enable estimation of community pharmacy loyalty, in-
dividuals had to have at least three pharmacy visits during the 365
days after antipsychotics initiation. This last criterion was needed to
obtain a reliable measure of an individual's pharmacy visiting
pattern. This threshold has been used in previous studies on
physician25e28 and pharmacy20,21 visiting patterns.
2.4. Variables
2.4.1. Community pharmacy loyalty
Community pharmacy loyalty was estimated as the number of
different community pharmacies visited by each individual for any
medication in the one-year period starting on the date of antipsy-
chotics initiation. In the province of Quebec, medication is usually
supplied for 30 days. Level of community pharmacy loyalty was
calculated for any medicationdnot only antipsychoticsdsince any
visit to the pharmacy, no matter the drug dispensed, is an oppor-
tunity for detecting non-adherence and providing information and
support. This global approach was also used in all the previous
studies on community pharmacy loyalty.18e21
55
2.4.2. Persistence and implementation
Distinct measures of persistence and implementation were
used to clearly distinguish these two different medication-taking
behaviours. Persistence was assessed using the treatment anni-
versary method.4 Individuals who were dispensed any antipsy-
chotic (not necessarily the one initially prescribed) before the
730th day following treatment initiation yet their days' supply
overlapped the 730th day were considered persistent. As in-
dividuals may be persistent but having suboptimal implementa-
tion, a permissible gap of 0.5 was added to the number of
overlapping days' supply. Implementation was estimated among
persistent individuals using the proportion of days covered
(PDC)29 by an antipsychotic (any antipsychotic). PDC was calcu-
lated for the time interval between the 365th day and 730th day
after antipsychotics initiation. Days of hospitalization were not
included in the PDC calculation because drugs received during
hospitalization are not recorded in the hospitalization database.
Individuals with PDC <0.8 were considered as having suboptimal
implementation based on a study showing that individuals with
schizophrenia who had a PDC <0.8 were more likely to be
admitted to hospital and to be hospitalized longer than patients
with PDC
0.8.8
2.4.3. Potential confounders
The potential confounders considered were patient charac-
teristics, drug treatment, and healthcare use factors that were
found to be associated with community pharmacy loyalty in our
previous study on use of multiple pharmacies among individuals
with schizophrenia22 and those shown to be associated with
persistence and/or implementation in other studies and that were
available in our databases.5,6,30,31 Patient characteristics at anti-
psychotics initiation included sex, age (20e29, 30e64,
65 years),
residency area (urban, rural), the GIS status, a proxy for socio-
economic status (yes, no), and history of a substance-use disor-
der (ICD-9 codes: 291-292; 303-305; ICD-10: F10-F19; F55) in the
365 days after antipsychotics initiation. Drug treatment-related
characteristics included the prescriber's specialty at antipsy-
chotics initiation (psychiatrist, general practitioner, or other), the
antipsychotic initially prescribed (typical antipsychotics, risperi-
done, olanzapine, quetiapine, clozapine), initial therapy regimen
(monotherapy, combined therapy), treatment intensity as the
proportion of the defined daily dose (DDD)32 of the initial anti-
psychotic measured the 30th day after its initiation (limited to
those having the initial antipsychotic in monotherapy; categories
based on tertiles < 40%, 40e100%, and >100%), use of a pill
organizer (which is defined as receiving 7-day supplies of anti-
psychotic medication for
80% of the 365 days after initiation), an
indicator for comorbidity as measured by the number of different
drugs used d based on International Non-proprietary Names
(INN) codesdin the 365 days after antipsychotics initiation (not
necessarily used simultaneously) (categories: 1e5, 6e10, 11e15,
16e20 and 21e51). This index has been found to be a good pre-
dictor of physician consultations, healthcare costs and mortal-
ity.33,34 Healthcare use factors included the number of pharmacy
claims in the 365 days after antipsychotics initiation and three
mutually exclusive profiles of healthcare use based on hospital
admissions reflecting the level of the severity and control of
schizophrenia: Profile 1 corresponded to individuals having only
ambulatory physician consultations for a mental health problem;
Profile 2 represented those having consultations for a mental
health problem with psychiatrists or other health professionals at
the emergency department; Profile 3 corresponded to those hos-
pitalized for a mental health problem; all other individuals were
assigned the category Other profile.35
56 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61
2.5. Statistical analyses
Descriptive statistics were used to describe patients' character-
istics and community pharmacy loyalty levels. The effect of com-
munity pharmacy loyalty level on persistence and implementation
was assessed separately. For each model, prevalence ratios (PR) and
95% confidence intervals (95%CI) were computed using a general-
ized linear model with a Poisson working model.36 Model-robust
variances were obtained with sandwich estimators to account for
the larger variance of Poisson variables compared to binomial
variables.37 First, univariate analyses were performed. To build
adjusted models, all potential confounding factors were entered in
the same multivariate model. Then, the influence of each factor on
the PR was verified by removing variables one by one. Only vari-
ables inducing a variation greater than 10% on the PR when
removed were kept in the final adjusted model. Linear trend across
levels of community pharmacy loyalty was tested using a linear
contrast for equally spaced categories.38 In each model, the po-
tential cluster effect of pharmacies was considered, considering the
pharmacy most frequently visited. This was done to control for the
fact that pharmacy characteristics and services offered could in-
fluence persistence and implementation among the group of in-
dividuals visiting the same pharmacy. The final models were tested
for collinearity using variance inflation factor tests.38,39 A
Fig. 2. Selection of study population.
sensitivity analysis was performed by including individuals having
<3 visits in a pharmacy.
2.6. Ethical considerations
This study was approved by the ethics board of the Centre hos-
pitalier universitaire de Qu

ebec, and the Commission d'acc  l'in-
es a
formation du Qu
ebec authorized access to the data.
3. Results
A total of 6,251 individuals with schizophrenia were included in
the study (Fig. 2). Their characteristics are presented in Table 1.
Among them, 52.3% were male and 65.0% were aged between 30
and 64 years. The majority (86.9%) were living in an urban area and
71.0% were recipients of GIS. Antipsychotics were prescribed by a
psychiatrist for 64.6% of the individuals with olanzapine being the
most frequently prescribed antipsychotic at initiation (34.3%).
Among those who had initiated antipsychotic treatment in mono-
therapy (n ¼ 5,633), 2,178 (38.7%) switched at least once to another
antipsychotic than the one initially prescribed over the 730 days of
the study (27.1% had at least one switch in the first 365 days and
33.1% had at least one switch between the 365th and the 730th day
after treatment initiation).
 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61 57

During the 365 days after antipsychotics initiation, 45.9%, 29.3%, after antipsychotics initiation, 33.8% of individuals were considered
13.8%, and 11% of individuals claimed their drugs in 1, 2, 3, and
4 non-persistent with antipsychotics and among them, 23.7% had
community pharmacies, respectively (Table 2). At the 730th day suboptimal implementation (PDC<80%) during the second year

Table 1
Characteristics of the 6,251 individuals with schizophrenia initiating antipsychotics.

Characteristics n ¼ 6,251

% (n)

Patient characteristics
Sex
Female 47.7 2,980
Male 52.3 3,271
Age at antipsychotics initiation (years):
20e29 19.3 1,209
30e64 65.0 4,065

65 15.6 977
Residency area at antipsychotics initiationa:
Urban 86.9 5,433
Rural 12.4 777
Unknown 0.7 41
Guaranteed income supplement (GIS) status:
No 29.0 1,812
Yes 71.0 4,439
Substance-use disorder in the year after antipsychotics initiationb:
No 83.8 5,237
Yes 16.2 1,014
Drug treatment-related characteristics
Time between schizophrenia diagnosis and antipsychotic initiation (days) (mean (±SD), median): 40 (±71) 12
Prescriber specialty
Psychiatrist 64.6 4,039
General practitioner or other 35.0 2,189
Unknown 0.4 23
Antipsychotics prescribed at initiation:
Typical 12.6 789
Atypical:
Risperidone 32.8 2,047
Olanzapine 34.3 2,141
Quetiapine 8.8 552
Clozapine 1.7 104
Combination of
2 antipsychotics 9.9 618
Proportion of the initial antipsychotic defined daily dose (DDD) the 30th day after antipsychotics initiation
< 40% 25.0 1,565
40e100% 28.4 1,778
> 100% 28.2 1,764
No treatment at the 30th day 8.4 526
Combination of
2 antipsychotics 9.9 618
Switch to another antipsychotic than the one initially prescribed among those having a monotherapy (n ¼ 5,633)
In the 365 days after initiation 27.1 1,524
Between the 365th and 730th day after initiation 33.1 1,864
Antipsychotics provided on a weekly-basis in the year after antipsychotics initiationc:
No 83.0 5,190
Yes 17.0 1,061
Number of different drugs used in the 365 days after antipsychotics initiationd:
1e5 24.0 1,503
6e10 10.9 678
11e15 27.0 1,687
16e20 17.0 1,062
21e51 21.1 1,321
Healthcare services use characteristics
Number of pharmacy claims in the 365 days after antipsychotics initiation:
3e12 24.0 1,498
13e24 36.2 2,266
25e368 39.8 2,487
Profiles of healthcare use in the 365 days after antipsychotics initiatione:
Profile 1: Only ambulatory physician consultations for mental health problem 48.7 3,044
Profile 2: Consultations for mental health problem with psychiatrists or other health professionals at the emergency department 14.2 890
Profile 3: Hospitalization for a mental health problem 28.9 1,805
Other profile 8.2 512

Total of percentage may not equal 100% because of rounding.

a
Determined according to the second position of the national postal code.
b
ICD-9 codes: 291-292; 303-305; ICD-10: F10-F19; F55.
c
80% of antipsychotics claimed for periods of seven days.
d
Based on INN codes.
e
Profiles are mutually exclusive. Individuals were categorized according to the more severe event. “Other profile” include individuals who did not have a consultation or a
hospitalization (n ¼ 88) and individuals who had a consultation and/or a hospitalization not related to a mental health problem (n ¼ 424).
58 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61
Table 2
Number of community pharmacies visited in the 365 days after antipsychotics
initiation (n ¼ 6,251).
n (%)
Number of community pharmacies visited
1 2,870
2 1,831
3 864

4 686
Mean (SD) a
2.0
Median 2.0
a
SD ¼ standard deviation.
after antipsychotics initiation. When compared to individuals who
had all their prescriptions filled at the same pharmacy, only in-
dividuals who had them filled in 3 or
4 different pharmacies were
more likely to be non-persistent (adjusted prevalence ratio (aPR)
1.17; 95% confidence intervals (CI) ¼ 1.06e1.29 and aPR ¼ 1.22; 95%
CI ¼ 1.10e1.37, respectively) (Table 3). When compared to in-
dividuals who had all their prescriptions filled at the same phar-
macy, those who had their prescriptions filled in
4 different
pharmacies were 49% more likely to have a PDC<80% (aPR ¼ 1.49;
95%IC ¼ 1.28e1.74). Linear trends for the prevalence of non-
persistence and suboptimal implementation across the number of
pharmacies were significant (p-value < 0.01). Similar results were
obtained for persistence and implementation from our sensitivity
analysis (Table 3a).
4. Discussion
In this cohort of individuals with schizophrenia who were new
antipsychotic users, visiting multiple pharmacies was frequent;
54.1% of individuals claimed their prescribed drugs in more than
one pharmacy over a one-year period. Individuals who visited four
different pharmacies or more (approximately 10% of the cohort)
were significantly more likely to discontinue their antipsychotics
within the two years after initiation and more likely to be exposed
to an antipsychotic for <80% of the days.
These results add to the emergent body of literature studying
the association between community pharmacy loyalty and
adherence,18e21 an important aspect of the quality of drug use. To
our knowledge, only one study specifically examined the associa-
tion between community pharmacy loyalty and persistence.20 In
this study, which was conducted by our team, individuals with
diabetes who visited a single pharmacy were more likely to be
persistent (OR 1.13, 95% CI 1.09e1.16). Three published studies
examined the association with implementation for different types
of physical illnesses,18,19,21 and use of a single pharmacy was posi-
tively associated with implementation. The first study was con-
ducted in the USA among individuals who were prescribed
cardiovascular medications.19 Results showed that for each addi-
tional pharmacy visited during the 3-month assessment window,
statin implementation ddefined as the proportion of days covered
d was reduced by 1.6% points. A study from the USA that included
926,956 individuals aged
65 years taking different types of drugs
for chronic illnesses indicated that those individuals who had
prescriptions filled at multiple pharmacies were 10%e31% more
likelyddepending on the drug studieddto be covered by a drug for
<80% of the days when compared to those who visited only one
pharmacy during a one-year period.18 In the study conducted by
our team among individuals with diabetes, those who visited a
single pharmacy were also more likely to have optimal imple-
mentation of their antidiabetic drugs (PDC
80%) (OR 1.22, 95% CI
1.19e1.26).20 In our study conducted among individuals with
diabetes as well as in the current study among individuals with
schizophrenia, the magnitude of the associations observed for
implementation is greater than for persistence. This may indicate
that the role of community pharmacists is likely to be more rele-
vant for refill monitoring and providing patients with strategies to
(45.9) integrate taking medication into their daily life and coping with
(29.3) side effects rather than preventing premature discontinuation.
(13.8)
This study is the first to assess the association between com-
(11.0)
(1.4) munity pharmacy loyalty and persistence and implementation
among individuals with a severe mental illness, namely schizo-
phrenia. Practice guidelines state that continuity of care is a core
general principle of schizophrenia management1,40,41 and our study
represents an important first step in exploring the relationship
between optimal drug use and more integrated care for antipsy-
chotic drug treatment in the community. Several factors may in-
fluence persistence and implementation of antipsychotics and a
large body of the literature has already focused on many patient-
related (e.g., socio-demographic, psychosocial) and treatment-
related (e.g., type of antipsychotic, doses) potential de-
terminants5,6,30,31,42,43 with less emphasis on organizational fac-
tors. Our results suggest that community pharmacy loyalty d a
healthcare system-related factor that is a core component of con-
tinuity of pharmaceutical caredis also associated to medication-
taking behaviour. Many hypotheses will need to be considered to
explain the mechanisms through which pharmacy loyalty may be
associated to adherence. For example, pharmacists have more
complete information on the medication of individuals who are
loyal to a single pharmacy and thus, may more easily detect and
intervene on adherence. By contrast, individuals that are less
adherent may prefer to claim their medications in different phar-
macies to avoid their non-adherence from being detected by
pharmacists.
This study has strengths. First, the cohort included new anti-
psychotic users that represented an important proportion of in-
dividuals with schizophrenia in the province of Quebec. Indeed, a
population-based study indicates that 75% of individuals with
schizophrenia in Quebec are covered by the public drug plan.44
However, individuals insured by the public drug plan may be
different from individuals insured through an employer plan and
results may not be generalizable to this latter group. Second, the
fact that adherence is a multidimensional construct4,45,46 was taken
into account by using the treatment anniversary method to assess
persistence and the PDC to assess implementation. These distinct
measures allowed us to distinguish between two different
medication-taking behaviours that are not necessarily driven by
the same factors,5 and that may require different interventions. In
addition, our method of taking into account any antipsychotic
claimed instead of being limited to the antipsychotic initially pre-
scribed prevented us from underestimating persistence and
implementation. Finally, the potential clustering effect of pharma-
cies on our results was assessed, an aspect that was only considered
in one study on pharmacy loyalty and adherence.21
This study also has some limitations. First, it adds new knowl-
edge specific to individuals with mental illness to the recent body of
research suggesting that more centralized pharmaceutical services
could contribute to improvement in pharmaceutical care,18,19 but
this hypothesis will need to be more formally tested in interven-
tional studies. Next, we wanted to include new antipsychotic users
only. However, it is possible that the 365-day look back period
selected to have an equal time-window for all individuals4 did not
allow for the exclusion of all individuals who had previously used
antipsychotics. The fact that past experience with antipsychotics
may have positively or negatively influenced medication persis-
tence or implementation can not be ruled out. Finally, information
on potential confounders such as psychosocial factors associated
 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61 59

Table 3
Prevalence ratios (PRs) comparing the proportion of individuals being non-persistent with antipsychotics and having suboptimal implementation according to the number of
community pharmacies visited.

Number of community pharmacies visited N Prevalence Crude Adjustedb

PRa (95% CI) Pc PR (95% CI) Pc

Non-persistence (n ¼ 6,251)
1 2,870 29.6 1.00 e 1.00 e
2 1,831 35.0 1.18 (1.07 1.07 (0.99
e1.30) e1.17)
3 864 39.5 1.33 (1.19 1.17 (1.06
e1.49) e1.29)

4 686 41.0 1.38 (1.22 1.22 (1.10
e1.56) e1.37)
<0.01 <0.01
Suboptimal implementationd (n ¼ 4,139)
1 2,020 20.3 1.00 e 1.00 e
2 1,191 23.6 1.17 (1.00 1.05 (0.92
e1.36) e1.19)
3 523 28.3 1.40 (1.18 1.15 (0.99
e1.66) e1.35)

4 405 35.3 1.74 (1.48 1.49 (1.28
e2.05) e1.74)
<0.01 <0.01
a
PR ¼ prevalence ratio; 95% CI ¼ 95% confidence intervals.
b
Adjusted models for non-persistence include: sex, guaranteed income supplement (GIS) status (yes, no) at antipsychotics initiation, prescriber specialty at antipsychotics
initiation (psychiatrist, general practitioner or other, unknown), proportion of the defined daily dose (DDD) of the initial antipsychotics the 30th day after first claim (for
monotherapy regimen only) (categories based on tertiles < 40%, 40e100% and >100%), having substance-use disorder, number of different drugs used in the 365 days after
antipsychotics initiation (categories: 1e5, 6e10, 11e15, 16e20 and 21e51 different types of drugs), claim of antipsychotics on a weekly-basis (yes, no), number of visits at
pharmacy, profiles of healthcare use on the basis of hospital admissions. Adjusted models for suboptimal implementation include the same variables with the exception of
prescriber's speciality and proportion of DDD of the antipsychotic dose.
c
P-value for linear trend test for ordinal variables obtained from a generalized linear model with a Poisson working model.
d
Suboptimal implementation was estimated among persistent only and defined has having <80% of the days covered by an antipsychotic.

with adherence to antipsychotics, including negative attitude to- 5. Conclusion

ward medication,47 poor insight,48 and quality of patient-
healthcare relationship31 which could also be associated with an-
tipsychotics adherence was not available. In addition, information
from the databases used did not allow determining if moving from
one neighbourhood to another also influences the use of multiple
pharmacies. However, the information available in our databases
was maximized by considering numerous patients characteristic,
drug treatment, and healthcare service related variables in our
adjusted analyses.
Results from our study have implications for healthcare teams
and researchers. These results could be used to sensitize healthcare
teams to the fact that individuals having very erratic pharmacy
visiting patterns are also more likely to be non-persistent and have
suboptimal implementation. Efforts should be made to identify
individuals whose pharmacy-visiting patterns are irregular and put
in place strategies to ensure adequate medication follow-up. In
addition, healthcare teams may present individuals with

Table 3a
Prevalence ratios (PRs) comparing the proportion of individuals being non-persistent with antipsychotics and having suboptimal implementation according to the number of
community pharmacies visited - Sensitivity analyses including individuals with
1 pharmacy visits.
b
Number of community pharmacies visited N Prevalence Crude Adjusted
a c
PR (95% CI) P PR (95% CI) Pc

Non-persistence (n ¼ 6,615)
1 3,177 34.7 1.00 e 1.00 e
2 1,888 36.2 1.04 (0.95e1.14) 1.02 (0.95e1.10)
3 864 39.5 1.14 (1.02e1.26) 1.10 (1.00e1.21)

4 686 41.0 1.18 (1.05e1.32) 1.16 (1.05e1.29)
<0.01 <0.01
Non-complianced (n ¼ 4,207)
1 2,074 21.6 1.00 e 1.00 e
2 1,205 24.2 1.12 (0.97e1.30) 1.04 (0.92e1.17)
3 523 28.3 1.31 (1.10e1.55) 1.13 (0.97e1.32)

4 405 35.3 1.63 (1.39e1.92) 1.48 (1.27e1.72)
<0.01 <0.01
a
PR ¼ prevalence ratio; 95% CI ¼ 95% confidence intervals.
b
Adjusted models for non-persistence include: sex, guaranteed income supplement (GIS) status (yes, no) at antipsychotics initiation, prescriber specialty at antipsychotics
initiation (psychiatrist, general practitioner or other, unknown), proportion of the defined daily dose (DDD) of the initial antipsychotics the 30th day after first claim (for
monotherapy regimen only) (categories based on tertiles < 40%, 40e100% and >100%), having substance-use disorder, number of different drugs used in the 365 days after
antipsychotics initiation (categories: 1e5, 6e10, 11e15, 16e20 and 21e51 different types of drugs), claim of antipsychotics on a weekly-basis (yes, no), number of visits at
pharmacy, profiles of healthcare use on the basis of hospital admissions. Adjusted models for implementation include the same variables with the exception of prescriber's
speciality and proportion of DDD of the antipsychotic dose.
c
P-value for linear trend test for ordinal variables obtained from a generalized linear model with a Poisson working model.
d
Suboptimal implementation was estimated among persistent only.
60 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61
schizophrenia the advantages of visiting the same pharmacy when
possible. Initiatives such as the Quebec computerized medical re-
cord (Dossier Sant
e Qu
ebec) that is being progressively imple-
mented and links health data from hospitals, pharmacies and
private clinics could facilitate the exchange of information on drug
claims if systematically used by all community pharmacists. In the
research perspective, this study and our previous work22 provided
a detailed description of pharmacy loyalty in the community
setting, one core component of continuity of pharmaceutical care.
Since 43% of the individuals in our cohort visited the emergency or
were hospitalized in the year after antipsychotics initiation, the
continuity of pharmaceutical care between the hospital and the
primary healthcare setting and its association with quality of drug
use should also be studied in the future.
Funding
This work was supported by the Laval University Chair on
Adherence to Treatments. The Chair on Adherence to Treatments
was funded through unrestricted grants from AstraZeneca Canada,
Merck Canada, Pfizer Canada, Sanofi Canada and the Prends soin de
toi program. At time of the study, F. Zongo had a post-doctoral
research scholar award from the Programme canadien de bourses
de la Francophonie; S. Lauzier has a research scholar award from the
Fonds de recherche du Qu
ebec e Sant
e in partnership with the
Institut national d'excellence en sant

e et en services sociaux (INESSS).
Authors' disclosures of potential conflicts of interest

goire and S. Lauzier, are researchers at the
J. Moisan, J.-P. Gre
Laval University Chair on Adherence to Treatments. The Chair on
Adherence to Treatments was funded through unrestricted grants
from AstraZeneca Canada, Merck Canada, Pfizer Canada, Sanofi
Canada and the Prends soin de toi program. Other authors have
indicated no potential conflicts of interest.
Previous presentation of the data
Brief highlights from this study were presented at the 27th
International Conference on Pharmacoepidemiology and Thera-
peutic Risk, Chicago, 14e17 August 2011 (Lauzier S, Moisan J,
Gre
goire J-P. Pharmacy Loyalty and Compliance with Antipsy-
chotics, oral podium presentation).
Acknowledgments
The authors are thankful to Eric Demers for the statistical
analysis and to Norma Perez for her contribution to bibliography
research.
References
1. Canadian Psychiatric Association. Clinical practice guidelines. Treatment of
schizophrenia. Can J Psychiatry. 2005;50:7Se57S.
2. World Health Organization. Adherence to Long-term Therapies: Evidence for
Action; 2003. Available from: http://www.who.int/chp/knowledge/
publications/adherence_report/en/. Accessed 14.12.15.
3. Vrijens B, De Geest S, Hughes DA, et al. A new taxonomy for describing and
defining adherence to medications. Br J Clin Pharmacol. 2012;73:691e705.
4. Gre
goire J-P, Moisan J. Assessment of Adherence to Drug Treatment in Database
Research. Drug Utilization Research: Methods and Applications. John Wiley &
Sons, Ltd; 2016:369e380.
5. Cooper D, Moisan J, Gregoire JP. Adherence to atypical antipsychotic treatment
among newly treated patients: a population-based study in schizophrenia.
J Clin Psychiatry. 2007;68:818e825.
6. Lacro JP, Dunn LB, Dolder CR, Leckband SG, Jeste DV. Prevalence of and risk
factors for medication nonadherence in patients with schizophrenia: a
comprehensive review of recent literature. J Clin Psychiatry. 2002;63:892e909.
7. Jonsdottir H, Opjordsmoen S, Birkenaes AB, et al. Predictors of medication
adherence in patients with schizophrenia and bipolar disorder. Acta Psychiatr
Scand. 2013;127:23e33.
8. Valenstein M, Copeland LA, Blow FC, et al. Pharmacy data identify poorly
adherent patients with schizophrenia at increased risk for admission. Med Care.
2002;40:630e639.
9. Velligan DI, Weiden PJ, Sajatovic M, et al. The expert consensus guideline se-
ries: adherence problems in patients with serious and persistent mental illness.
J Clin Psychiatry. 2009;70(Suppl 4):1e46.
10. Gilmer TP, Dolder CR, Lacro JP, et al. Adherence to treatment with antipsychotic
medication and health care costs among Medicaid beneficiaries with schizo-
phrenia. Am J Psychiatry. 2004;161:692e699.
11. Morken G, Widen JH, Grawe RW. Non-adherence to antipsychotic medication,
relapse and rehospitalisation in recent-onset schizophrenia. BMC Psychiatry.
2008;8:32.
12. Ascher-Svanum H, Faries DE, Zhu B, Ernst FR, Swartz MS, Swanson JW. Medi-
cation adherence and long-term functional outcomes in the treatment of
schizophrenia in usual care. J Clin Psychiatry. 2006;67:453e460.
13. Dibonaventura M, Gabriel S, Dupclay L, Gupta S, Kim E. A patient perspective of
the impact of medication side effects on adherence: results of a cross-sectional
nationwide survey of patients with schizophrenia. BMC Psychiatry. 2012;12:20.
14. Bell S, McLachlan AJ, Aslani P, Whitehead P, Chen TF. Community pharmacy
services to optimise the use of medications for mental illness: a systematic
review. Aust N. Z Health Policy. 2005;2:29.
15. Al-Jumah KA, Qureshi NA. Impact of pharmacist interventions on patient's
adherence to antidepressants and patient-reported outcomes: a systematic
review. Patient Prefer Adherence. 2012;6:87e100.
16. Buurma H, Bouvy ML, De Smet PA, Floor-Schreudering A, Leufkens HG,
Egberts AC. Prevalence and determinants of pharmacy shopping behaviour.
J Clin Pharm Ther. 2008;33:17e23.
17. Pottegard A, Hallas J. Physicians' and pharmacies' overview of patients'
medication. An analysis of fidelity coefficients. Eur J Clin Pharmacol. 2011;67:
919e924.
18. Marcum ZA, Driessen J, Thorpe CT, Gellad WF, Donohue JM. Effect of multiple
pharmacy use on medication adherence and drug-drug interactions in older
adults with Medicare Part D. J Am Geriatr Soc. 2014;62:244e252.
19. Choudhry NK, Fischer MA, Avorn J, et al. The implications of therapeutic
complexity on adherence to cardiovascular medications. Arch Intern Med.
2011;171:814e822.
20. Dossa R, Gre
goire J-P, Lauzier S, Gue

nette L, Sirois C, Moisan J. Effect of loyalty
to a pharmacy on antidiabetes drug adherence and use of guidelines-
recommended drugs. Can J Diabetes. 2014;38:S71.
21. Russell C. Does the Continuity of Pharmaceutical Care Influences Adherence to
Statins?. Halifax, Nova Scotia: Dalhousie University; 2013. Master of sciences
http://dalspace.library.dal.ca/bitstream/handle/10222/42712/Christie-Russell-
MSc-CHE-September-2013.pdf?sequence¼7. Accessed 14.11.20.
22. Lauzier S, Gregoire JP, Lesage A, Moisan J. Community pharmacy loyalty among
individuals with schizophrenia. Res Soc Adm Pharm. 2013;9:806e816.
23. Re
gie de l'assurance maladie du Que
bec. Rapport annuel de gestion 2013-2014;
2014. http://www.ramq.gouv.qc.ca/fr/regie/salle-presse/actualites/2014/Pages/
rapport-annuel-2013-2014.aspx. Updated September 29, 2014. Accessed
17.12.10.
24. Dalman C, Broms J, Cullberg J, Allebeck P. Young cases of schizophrenia iden-
tified in a national inpatient registereare the diagnoses valid? Soc Psychiatry
Psychiatr Epidemiol. 2002;37:527e531.
25. Chen CC, Chen SH. Better continuity of care reduces costs for diabetic patients.
Am J Manag Care. 2011;17:420e427.
26. Lin W, Huang IC, Wang SL, Yang MC, Yaung CL. Continuity of diabetes care is
associated with avoidable hospitalizations: evidence from Taiwan's National
Health Insurance scheme. Int J Qual Health Care. 2010;22:3e8.
27. Gill JM, Mainous 3rd AG, Nsereko M. The effect of continuity of care on
emergency department use. Arch Fam Med. 2000;9:333e338.
28. Brousseau DC, Meurer JR, Isenberg ML, Kuhn EM, Gorelick MH. Association
between infant continuity of care and pediatric emergency department utili-
zation. Pediatrics. 2004;113:738e741.
29. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy
records: methods, validity, and applications. J Clin Epidemiol. 1997;50:
105e116.
30. Higashi K, Medic G, Littlewood KJ, Diez T, Granstrom O, De Hert M. Medication
adherence in schizophrenia: factors influencing adherence and consequences
of nonadherence, a systematic literature review. Ther Adv Psychopharmacol.
2013;3:200e218.
31. Sendt KV, Tracy DK, Bhattacharyya S. A systematic review of factors influencing
adherence to antipsychotic medication in schizophrenia-spectrum disorders.
Psychiatry Res. 2015;225:14e30.
32. World Health Organization. International language for drug utilization
research. Available from:: http://www.whocc.no/ddd/application_form/.
updated December 10, 2014. Accessed 14.12.15.
33. Von Korff M, Wagner EH, Saunders K. A chronic disease score from automated
pharmacy data. J Clin Epidemiol. 1992;45:197e203.
34. Schneeweiss S, Seeger JD, Maclure M, Wang PS, Avorn J, Glynn RJ. Performance
of comorbidity scores to control for confounding in epidemiologic studies us-
ing claims data. Am J Epidemiol. 2001;154:854e864.
35. Benigeri M. L’utilisation des services de sant
e mentale par les Montr
ealais en
2004-2005. Montr
eal: Carrefour montr
ealais d’information sociosanitaire (CMIS)
de la Direction des ressources humaines. Information et planification de l’Agence
 F.E. Zongo et al. / Research in Social and Administrative Pharmacy 14 (2018) 53e61

et des services sociaux de Montre
de la sante
al; 2007.
36. Lumley T, Kronmal R, Ma S. Relative Risk Regression in Medical Research: Models,
Contrasts, Estimators, and Algorithms. UW Biostatistics Working Paper Series;
2006. Working Paper 293 http://biostats.bepress.com/uwbiostat/paper293.
Accessed 14.12.10.
37. Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios
and differences. Am J Epidemiol. 2005;162:199e200.
38. Kleinbaum D, Kupper L, Nizam A, Muller K. Applied Regression Analysis and
Other Multivariate Methods. third ed. 1998.
39. Marquardt DW. Generalized inverses, ridge regression biased linear estimation
and non-linear estimation. Technometrics. 1970;12:591e612.
40. Royal Australian and New Zealand College of Psychiatrists Clinical Practice
Guidelines Team for the Treatment of Schizophrenia and Related Disorders.
Royal Australian and New Zealand College of Psychiatrists clinical practice
guidelines for the treatment of schizophrenia and related disorders. Aust N. Z J
Psychiatry. 2005;39:1e30.
41. National institute for health and care excellence (NICE). Psychosis and Schizo-
phrenia in Adults: Treatment and Management; 2014. http://www.nice.org.uk/
guidance/cg178/resources/guidance-psychosis-and-schizophrenia-in-adults-
treatment-and-management-pdf. Accessed 15.01.10.
61
42. Agarwal MR, Sharma VK, Kishore Kumar KV, Lowe D. Non-compliance with
treatment in patients suffering from schizophrenia: a study to evaluate
possible contributing factors. Int J Soc Psychiatry. 1998;44:92e106.
43. Acosta FJ, Hernandez JL, Pereira J, Herrera J, Rodriguez CJ. Medication adher-
ence in schizophrenia. World J Psychiatry. 2012;2:74e82.
44. Vanasse A, Courteau J, Fleury MJ, Gre
goire JP, Lesage A, Moisan J. Treatment
prevalence and incidence of schizophrenia in Quebec using a population health
services perspective: different algorithms, different estimates. Soc Psychiatry
Psychiatr Epidemiol. 2011;47:533e543.
45. Cramer JA, Roy A, Burrell A, et al. Medication compliance and persistence:
terminology and definitions. Value Health. 2008;11:44e47.
46. Wahl C, Gregoire JP, Teo K, et al. Concordance, compliance and adherence in
healthcare: closing gaps and improving outcomes. Healthc Q. 2005;8:65e70.
47. Cabeza IG, Amador MS, Lopez CA, Gonzalez de Chavez M. Subjective response
to antipsychotics in schizophrenic patients: clinical implications and related
factors. Schizophr Res. 2000;41:349e355.
48. Olfson M, Mechanic D, Hansell S, Boyer CA, Walkup J, Weiden PJ. Predicting
medication noncompliance after hospital discharge among patients with
schizophrenia. Psychiatr Serv. 2000;51:216e222.