SCIENCE AND PRACTICE

Journal of the American Pharmacists Association 58 (2018) S78eS82

Contents lists available at ScienceDirect

Journal of the American Pharmacists Association

journal homepage: www.japha.org

RESEARCH NOTES
Relationship between medication synchronization and
antiretroviral adherence
Emily Ghassemi*, Jennifer Smith, Laura Owens, Charles Herring, Melissa Holland

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

Article history: Objectives: To compare antiretroviral adherence (measured as the proportion of days covered
Received 6 August 2017 [PDC]) and change in viral load in insured, HIV-infected, adult outpatients enrolled and not
Accepted 7 May 2018 enrolled in a medication synchronization program.
Available online 12 June 2018
Methods: This was a multicenter, retrospective, pilot cohort study. Fifty-eight insured,
HIV-infected, outpatients at least 18 years of age receiving antiretroviral therapy (ART) for at
least 3 months as of August 2015 were included. PDC, viral load, PDC dichotomized into
adherent or nonadherent, and viral load dichotomized into detectable or undetectable were
collected for each patient. Study data were compared in those with (enrolled) and without (not
enrolled or control) medication synchronization. The study end points were analyzed between
the 2 groups retrospectively after 3 months.
Results: PDC in patients undergoing medication synchronization was significantly higher than
in control patients: mean ± SD 96 ± 9% versus 71 ± 27%, respectively (P < 0.0001). The
medication synchronization group was also more likely to be adherent to ART than the control
group (odds ratio 10.67, 95% confidence interval 2.63e43.31). In the medication synchroni-
zation group, 75.9% of patients had an undetectable baseline viral load, and 83.3% had an
undetectable viral load at study completion. In the control group, 62.1% and 64.7% had an
undetectable viral load at baseline and completion, respectively. No statistically significant
change in viral load was observed between groups (P ¼ 0.34).
Conclusion: In insured, HIV-infected, adult outpatients, implementation of a medication
synchronization program was associated with improved ART adherence. Future studies are
needed to better assess the impact of medication synchronization on clinical outcomes.
© 2018 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Medication adherence to antiretroviral therapy (ART) is
important in patients living with human immunodeficiency
virus (HIV)/acquired immunodeficiency syndrome (AIDS) to
achieve viral suppression, minimize drug resistance, and
improve survival.1 The Department of Health and Human
Services Panel on Antiretroviral Guidelines identifies many
factors that may influence adherence. These factors include
the prescribed regimen, health literacy, substance abuse,
Disclosure: The authors declare no conflicts of interest, real or apparent, and
no financial interests in any company, product, or service mentioned in this
program, including grants, employment, gifts, stock holdings, and honoraria.
Previous presentations: Poster presentation at the 2016 American Pharma-
cists Association Annual Meeting, Baltimore, MD; Podium presentation at the
2016 Research in Education and Practice Symposium, UNC Eshelman School
of Pharmacy, Chapel Hill, NC.
* Correspondence: Emily Ghassemi, 303 Green Street East, Wilson,
NC 27893.
E-mail address: ekfretz0529@email.campbell.edu (E. Ghassemi).
poverty, inconsistent access to medication, and psychosocial
hurdles (such as depression or HIV stigma). A supportive
clinical setting with multidisciplinary care, including pro-
viders, pharmacists, and behavioral health clinicians, was
associated with higher adherence.2
ART can be costly, which may affect a patient’s access to
medication. Uninsured HIV-infected adults may be eligible to
receive ART and other associated medications at no cost
through the nationwide AIDS Drug Assistance Program
(ADAP).3 Insured patients with HIV/AIDS are not eligible to
receive their medications through ADAP or similar programs,
which could affect medication adherence.
A number of surrogate endpoints have been used to mea-
sure medication adherence in previous studies, including refill
adherence, pill counts, electronic monitoring devices, medi-
cation possession ratio, and proportion of days covered (PDC).
The PDC calculates the number of days supplied of a medica-
tion class (such as ARTs) obtained over a specified time
period.4 A variety of techniques and interventions have been

https://doi.org/10.1016/j.japh.2018.05.002
1544-3191/© 2018 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Medication synchronization and adherence
proven to affect adherence. These include compliance pack-
aging and educational interventions, such as counseling.5
More recently, medication synchronization has been used in
areas such as chain, independent, and mail-order pharmacy,
most often targeting polypharmacy patients and the
nonadherent.
Medication synchronization is the process of refilling all
maintenance medications on the same date for each fill cycle
and is intended to improve medication adherence as well as
pharmacy operation efficiency.6 In a traditional model, the
pharmacy staff prepares medications only when prompted, for
example, by a patient requesting refills or a provider ordering a
new prescription. In medication synchronization, a date is
deliberately scheduled for the pharmacy staff to fill all of a
patient’s chronic medications. This model is advantageous for
pharmacy workflow. By scheduling and synchronizing
medications, pharmacy staff are able to perform medication
reviews, identify any potential therapeutic concerns, and
complete patient counseling on all medications during 1
scheduled date.7 In addition, the synchronization model is
advantageous for patients by reducing trips to the pharmacy
and allowing 1 convenient day to pick up all medications.
Medication synchronization and improved adherence may
also have a cost benefit with reduced disease-related medical
costs.8 Holdford et al. published 2 studies to assess the impact
of appointment-based medication synchronization (ABMS)
and adherence to chronic medications in the general popula-
tion. Both studies found a higher odds of adherence in the
ABMS groups compared with control groups.9,10 Data on
medication synchronization and adherence are limited, and to
date no published studies have investigated medication
synchronization for ART. Therefore, the purpose of the present
study was to determine if enrollment in medication synchro-
nization was associated with improved adherence to ART.
Objectives
The primary objective of this investigation was to compare
ART adherence (as measured by PDC) in insured, HIV-infected,
adult outpatients enrolled and not enrolled in medication
synchronization. The secondary objective was to describe viral
load in both medication synchronization and control groups
before and after medication synchronization.
Methods
Study design and period
This was a multicenter, retrospective, pilot cohort study.
The study period was August 2015 to April 2016.
Setting and medication synchronization program
The study site was a Federally Qualified Health Center in
eastern North Carolina. Patients were identified and divided
into “medication synchronization” (enrolled into the medica-
tion synchronization program) and “control” (not yet enrolled
in the medication synchronization program). Because of the
retrospective, observational nature of this study, the groups
were predetermined by the pharmacy’s ongoing medication
synchronization implementation and no matching or
SCIENCE AND PRACTICE
randomization was performed by the investigators. The
medication synchronization group differed from the control
group in timing of refills for chronic medications. When
patients in the control group required a refill for a chronic
medication, the patient would prompt the pharmacy to refill
each medication individually at the time it was due, whereas
all chronic medications were refilled on the same date for
patients in the medication synchronization group.
Patient eligibility
Patients were eligible for inclusion in the study if they were
outpatients at least 18 years of age, diagnosed with HIV (based
on medical records) as of August 2015, using any one of the 3
Carolina Family Health Centers, Inc. pharmacies, prescribed a
consistent ART regimen for at least 3 months, and insured,
either publicly or privately. ART regimen was defined as the
combination of antiretroviral medications prescribed for the
treatment of HIV. Patients were excluded for any missing or
incomplete data for calculation of the primary end point (PDC)
and for any ART regimen change during the study period.
Patients were identified for study eligibility by a “Ryan White”
group identification code in the pharmacy dispensing software
system. All patients eligible for enrollment in the medication
synchronization program at data collection were eligible for
inclusion in the study (enrolled group). A group of individuals
meeting the same inclusion criteria, but not enrolled in the
medication synchronization program, were randomly selected
and identified as the control group and received no
intervention.
Outcome measures
The primary end point (PDC) was determined for each
patient with the use of the following equation: PDC equals the
number of days supplied in the given period divided by the
number of days in the given period times 100%. Though not
validated, PDC is a common measure of adherence in the
literature. After its determination, the PDC is often dichoto-
mized as adherent or nonadherent. Andrade et al. found 38
studies in which PDC was categorized into levels of adherence.
Of the studies that dichotomized PDC, adherence was deter-
mined by a PDC of 80% or higher in 24 (75%) of the studies and
90% or higher in 4 (13%).11 Consistent with previous studies,
the present investigation defined adherence by a PDC of 80% or
higher. ART and PDC data were collected from the electronic
medical record and pharmacy dispensing system software.
Viral load data were collected from the electronic medical
record. If available, 2 viral loads were recorded for each pa-
tient, a preemedication synchronization viral load, which was
drawn before the start of the synchronization program, and a
postemedication synchronization viral load, which was drawn
3 months after the start of the program. Viral load was
considered to be undetectable at fewer than 20 copies/mL.
Other independent variables
Three months of data were collected retrospectively from
the electronic medical record. Data collected included de-
mographic data (age, gender, ethnicity, and race) and clinical
data (comorbidities and medications, including single-drug vs.
S79
SCIENCE AND PRACTICE
E. Ghassemi et al. / Journal of the American Pharmacists Association 58 (2018) S78eS82

multidrug ART regimens). Refill dates and days supplied were Table 1
recorded for each patient to calculate PDC. Automatic refills Baseline characteristics, n (%)

were not used, because patients had to prompt the refills. All Characteristic Medication Control
study end points were collected 3 months after the medication synchronization (n ¼ 29) (n ¼ 29)
synchronization group was enrolled into the program. Male 20 (69) 19 (66)
Age, y, mean ± SD 52 ± 10 45 ± 11
Race
Statistical analysis African American 27 (93) 26 (90)
White 1 (3) 3 (10)
Approximately 36 to 42 patients per group were needed to Other 1 (3) 0 (0)
show a significant difference of 0.15-0.25 (effect size based on Payer type
Medicare 17 (59) 14 (48)
previous studies) in the primary endpoint at a preset alpha
Medicaid 7 (24) 11 (38)
error of 5% and power of 80% to 85%.12 The mean PDC for each Private insurance 5 (17) 2 (14)
group was compared with the use of a 2-sample t test. For the Distance from primary residence to pharmacy
secondary endpoints, the change between starting and 0e5 miles 18 (62) 13 (45)
concluding viral loads was determined for each patient and 6e19 miles 7 (24) 7 (24)
compared between groups with the use of a 2-sample t test. 20e55 miles 3 (10) 7 (24)
Viral load and PDC were dichotomized and analyzed cate- Unlisted address 1 (3) 2 (7)
Number of patients with comorbidities
gorically. Univariate chi-square tests or Fisher exact tests
Hypertension 14 (48) 13 (45)
were used for the categoric endpoints of adherence and viral Dyslipidemia 8 (28) 7 (24)
load; odds ratio (OR), corresponding 95% confidence interval, Diabetes 6 (21) 4 (14)
and P values were reported. Due to the retrospective, Hepatitis 5 (17) 3 (10)
hypothesis-generating nature of this study, no adjustments Depression 4 (14) 9 (31)
for multiple comparisons were made. A P value of < 0.05 was Pulmonary disorder 8 (28) 3 (10)
Average number of chronic medications prescribed
considered to be significant for all analyses. Statistical ana-
1e4 3 (10) 10 (35)
lyses were performed with the use of JMP (version 10). The 5e10 19 (66) 12 (41)
study was approved by the Campbell University Institutional 11e23 7 (24) 7 (24)
Review Board. ART regimen
Single-drug 14 (48) 17 (59)
Multidrug 15 (52) 12 (41)
Results
Abbreviation used: ART, antiretroviral therapy.

Fifty-eight patients met the criteria and were included in

the study. Of the 32 patients enrolled in the medication
synchronization program, 29 were eligible for inclusion in the
study; 2 patients were not eligible for inclusion because of
changes to their ART regimen during the study period, and 1
patient was ineligible because he or she did not have an HIV
diagnosis. An additional 29 patients not enrolled in the
medication synchronization program, but eligible for inclusion
in the study, were randomly selected as the control group.
Baseline characteristics of the study population are presented
in Table 1. The mean age was 48.6 ± 11 years; male and
African-American patients accounted for 67% and 91% of the
study population, respectively. Patients taking a single-drug
ART regimen accounted for 53% of the study population.
Results of the primary and secondary analyses are sum-
marized in Table 2. Mean PDCs were 96 ± 9% and 71 ± 21% in
the medication synchronization and control groups, respec-
tively (P < 0.001). The medication synchronization group was
significantly more likely to be adherent than the control group
(OR 10.67, 95% CI 2.63e43.31). At baseline, 75.9% and 62.1%
had undetectable viral load in the medication synchronization
and control groups, respectively. Due to limitations in patient
follow-up (e.g., missed appointments, financial restraints for
labs), a postemedication synchronization viral load was
available for only 12 patients in the medication synchroniza-
tion group and 17 patients in the control group. At study
completion, 83.3% and 64.7% had an undetectable viral load in
medication synchronization and control groups, respectively.
The mean difference from baseline to end-of-study viral loads
between medication synchronization and control groups
(3.3log10 and 3.8log10, respectively; P ¼ 0.34) was not
significant. No significant difference was found for undetect-
able viral load between the medication synchronization and
control groups (P ¼ 0.27).
Discussion
In this study, medication synchronization was associated
with improved ART adherence. The population enrolled in the
medication synchronization program was almost 11 times
more likely to be adherent than those not enrolled in medi-
cation synchronization. These results are consistent with
previous studies on medication synchronization and adher-
ence in patient populations with other chronic disease states.
In contrast, no significant difference was noted between
groups in viral load.
In addition to improving adherence, previous research has
investigated patient satisfaction with medication synchroni-
zation programs.6 Although the present investigation did not
assess patient satisfaction with the program, improved
adherence in the medication synchronization group studied
was consistent with the patient-reported results of previous
studies.
Earlier research suggested that adherence rates of 95% or
higher were necessary to achieve virologic suppression in
HIV-infected individuals.13,14 Since the introduction of newer,
boosted ART regimens, Gordon et al. explored the relation-
ship between ART regimen adherence and virologic failure.
ART adherence of 80% to 90% was associated with a 3.5%
failure rate, whereas adherence of 90% or more was

S80

Medication synchronization and adherence
Table 2
Association between medication synchronization and PDC and viral load
Variable Medication
synchronization (n ¼ 29)
Mean PDC, % (95% CI) 96 (92e99)
Mean difference in viral load, log10 (95% CI) 3.3 (3.3 to 3.8)
Preintervention viral load, n (%)
<20 copies/mL 22 (55.0)
20 copies/mL 7 (38.9)
Postintervention viral load, n (%)
<20 copies/mL 10 (47.2)
20 copies/mL 2 (25.0)
Adherent, n (%) 26 (89.7)
Nonadherent, n (%) 3 (10.3)
P values, odds ratios, and corresponding 95% CIs were not adjusted for other factors.
Abbreviations used: CI, confidence interval; OR, odds ratio; PDC, proportion of days covered.
a
Two-sample t test (independent groups).
b
Chi-square test of association.
c
Fisher exact test of association.
associated with a 1.1% failure rate.1 ART resistance has been
associated with adherence less than 80%.15 Although viral
suppression may be achieved with ART adherence rates lower
than 95%, high adherence should remain the goal for treat-
ment success. With a mean PDC of 96 ± 9%, the medication
synchronization group in this study achieved even the most
rigorous adherence goal of greater than 95% referenced in
previous studies.
Previous research assessing the impact of medication
synchronization programs has not included HIV-infected pa-
tients and evaluated medication classes other than ART.9,10,16
The results of the present investigation align with those of
previous studies associating medication synchronization and
improved PDC in drug classes for other chronic disease states.
The results of the univariate analysis in this investigation
found even greater odds of adherence to ART in patients with
HIV/AIDS enrolled in medication synchronization compared
with other chronic disease states that have been studied: 10.67
times compared with 6.1 times.9
This study had several limitations. The study was initiated
while the pharmacies were in the process of introducing a
medication synchronization program, and the full impact of
the program may not have been evident owing to the limited
number of patients enrolled and length of the investigation.
The lack of demographic diversity in the study population
limits the external validity of the results; however, the age,
race, and gender of the study population strongly aligns with
the prevalence statistics of persons diagnosed with HIV in the
United States.17 Additional baseline data, such as baseline
adherence, time since HIV diagnosis, and time since starting
ART, were not collected and could affect results. Viral load
suppression generally takes 8 to 24 weeks in adherent
patients and in some instances may take longer.2 A longer
study period, larger sample size, or the availability of more
postintervention viral load results may have detected a
greater difference in viral load between the 2 groups. The
observational nature and sample size prohibited assessment
of the impact of demographic predictors of adherence, such
as age, gender, race, ethnicity, type of insurance, and type of
ART regimen (i.e., single dose versus multidose). Additional
factors could have affected adherence, including distance
between patient residence and pharmacy or patient method
SCIENCE AND PRACTICE
Control (n ¼ 29) Univariate P value
OR (95% CI)
71 (61e82) < 0.001a
3.8 (3.3 to 4.2) 0.34a
0.52 (0.17e1.62) 0.26b
18 (45.0)
11 (61.1)
0.37 (0.06e2.25) 0.41c
11 (52.4)
6 (75.0)
13 (44.8) 10.67 (2.63e43.31) < 0.001b
16 (55.2)
of obtaining medication (e.g., delivery); however, all patients
were offered the same options for pharmacy services,
including delivery. The retrospective nature of this pilot
study limits the results to hypothesis-generating. In addition,
the PDC is a surrogate marker of adherence. A PDC of 80%
indicates that a patient refilled sufficient medication to cover
80% of the days in the study period but does not necessarily
indicate that the medication was administered. In addition, a
higher threshold of adherence (i.e., 90%) would perhaps have
been more appropriate for the HIV population. Prospective
investigations should focus on clinical indicators of adher-
ence and outcomes with medication synchronization in
HIV-infected patients, such as viral load. Finally, 2 patients in
the medication synchronization group received medications
dispensed in a pillbox, which could have affected their
adherence.
Conclusion
Enrollment in a medication synchronization program
significantly improved ART adherence in a population of
adult HIV-infected outpatients. These results suggest that
the implementation of a medication synchronization pro-
gram by pharmacies serving such patients would be a
reasonable strategy to improve adherence to ART. Future
research should be extended over a longer study period and
include the impact of medication synchronization on clinical
endpoints, quality of life, and morbidity and mortality
outcomes.
Acknowledgments
The authors thank the Campbell University College of
Pharmacy and Health Sciences Pharmacy Practice Research
Committee, Michael Jiroutek, DrPH, MS, Associate Professor,
Department of Clinical Research, Campbell University College
of Pharmacy and Health Sciences, and JAPhA mentor, Patrick
Clay, PharmD, AAHIVP, CPI, CCTI, FCCP, Professor of Pharma-
cotherapy, University of North Texas System College of
Pharmacy.
S81
SCIENCE AND PRACTICE
E. Ghassemi et al. / Journal of the American Pharmacists Association 58 (2018) S78eS82

References 11. Andrade SE, Kahler KH, Frech F, Chan KA. Methods for evaluation of
medication adherence and persistence using automated databases.
Pharmacoepidemiol Drug Saf. 2006;15(8):565e574.
1. Gordon LL, Gharibian D, Chong K, Chun H. Comparison of HIV viro-
12. DSS Research Group. DSS research calculator. Available at: https://www.
logic failure rates between patients with variable adherence of three
dssresearch.com/KnowledgeCenter/toolkitcalculators/samplesizecalcu
antiretroviral regimen types. AIDS Patient Care STDS. 2015;29(7):
lators.asp. Accessed August 19, 2015.
384e388.
13. Shuter J. Forgiveness of non-adherence to HIV-1 antiretroviral therapy.
2. Department of Health and Human Services Panel on Antiretroviral
J Antimicrob Chemother. 2008;61(4):769e773.
Guidelines for Adults and Adolescents. Guidelines for the use of antire-
14. Paterson DL, Swindells S, Mohr J, et al. Adherence to protease inhibitor
troviral agents in HIV-1einfected adults and adolescents. Published April
therapy and outcomes in patients with HIV infection. Ann Intern Med.
8, 2015; updated January 28, 2016. Available at: http://aidsinfo.nih.gov/
2000;133(1):21e30.
contentfiles/lvguidelines/AdultandAdolescentGL.pdf. Accessed February
15. Meresse M, March L, Kouanfack C, et al. Patterns of adherence to anti-
21, 2016.
retroviral therapy and HIV drug resistance over time in Stratall ANRS
3. ADAP Advocacy Association. North Carolina ADAP Directory. Available at:
12110/ESTHER trial in Cameroon. HIV Med. 2014;15(8):478e487.
http://adap.directory/north-carolina#field_formulary_link. Accessed
16. Nguyen E, Sobieraj DM. The impact of appointment-based medication
August 21, 2015.
synchronization on medication taking behaviour and health outcomes: a
4. Choudhry NK, Shrank WH, Levin RL, et al. Measuring concurrent adher-
systematic review. J Clin Pharm Ther. 2017;42(4):404e413.
ence to multiple related medications. Am J Manag Care. 2009;15(7):
17. Centers for Disease Control and Prevention. HIV surveillance report,
457e464.
2016, vol. 28. November 2017. Available at: https://www.cdc.gov/hiv/
5. Peterson AM, Takiya L, Finely R. Meta-analysis of trials of interventions to
pdf/library/reports/surveillance/cdc-hiv-surveillance-report-2016-vol-28.
improve medication adherence. Am J Health Syst Pharm. 2003;60(7):
pdf. Accessed January 1, 2018.
657e665.
6. National Community Pharmacists Association. National Adherence
Emily Ghassemi, PharmD, MSCR, BCACP, CPP, Clinical Pharmacist, Carolina
Survey: the promise and prospects of medication synchronization.
Family Health Centers, Inc., Wilson, NC; at time of study: Pharmacy Resident,
August 2014. Available at: http://www.ncpa.co/pdf/med-sync-national-
Campbell University College of Pharmacy & Health Sciences, and Wilson Com-
adherence-survey.pdf. Accessed July 10, 2015.
munity Health Center, Wilson, NC
7. American Pharmacists Association Foundation. Pharmacy’s appointment
based model: implementation guide for pharmacy practices. July 2013. Jennifer Smith, PharmD, BC-ADM, BCACP, CDE, Clinical Pharmacy Specialist,
Available at: http://www.aphafoundation.org/sites/default/files/ckeditor/ U.S. Department of Veterans Affairs, Anderson, SC; at time of study: Associate
files/ABMImplementationGuide-FINAL-20130923.pdf. Accessed July 10, Professor of Pharmacy Practice, Campbell University College of Pharmacy &
2015. Health Sciences, and Wilson Community Health Center, Wilson, NC
8. Patterson JA, Holdford DA, Saxena K. Cost-benefit of appointment-based
medication synchronization in community pharmacies. Am J Manag Care. Laura Owens, PharmD, AAHIVP, Chief Pharmacy Officer, Carolina Family Health
2016;22(9):587e593. Centers, Inc., Wilson, NC
9. Holdford DA, Inocencio TJ. Adherence and persistence associated with an Charles Herring, BSPharm, PharmD, CPP, Associate Professor of Pharmacy
appointment-based medication synchronization program. J Am Pharm Practice, Campbell University College of Pharmacy & Health Sciences, and
Assoc. 2013;53(6):576e583. Downtown Health Plaza, Wake Forest Baptist Health, Winston-Salem, NC
10. Holdford D, Saxena K. Impact of appointment-based medication
synchronization on existing users of chronic medications. J Manag Care Melissa Holland, PharmD, MSCR, Assistant Professor of Clinical Research,
Pharm. 2015;21(8):662e669. Campbell University College of Pharmacy & Health Sciences, Buies Creek, NC

S82