Pharmacist Interventions to Improve Cardiovascular Disease ...MedicalSubjectHeadings(MeSH)andtext terms with Boolean operators. The syntax consisted of three search themes inter-sected

Pharmacist Interventions to ImproveCardiovascular Disease Risk Factors inDiabetesA systematic review and meta-analysis of randomized controlled trials

VALÉRIE SANTSCHI, PHARMD, PHD1,2

ARNAUD CHIOLERO, MD, MSC, PHD1,2

GILLES PARADIS, MD, MSC1,3,4

APRIL L. COLOSIMO, MSC, MLIS5

BERNARD BURNAND, MD, MPH2

OBJECTIVEdThis systematic review and meta-analysis of randomized controlled trials(RCTs) assesses the effect of pharmacist care on cardiovascular disease (CVD) risk factors amongoutpatients with diabetes.

RESEARCH DESIGN AND METHODSdMEDLINE, EMBASE, CINAHL, and theCochrane Central Register of Controlled Trials were searched. Pharmacist interventions wereclassified, and a meta-analysis of mean changes of blood pressure (BP), total cholesterol (TC),LDL cholesterol, HDL cholesterol, and BMI was performed using random-effects models.

RESULTSdThe meta-analysis included 15 RCTs (9,111 outpatients) in which interventionswere conducted exclusively by pharmacists in 8 studies and in collaboration with physicians,nurses, dietitians, or physical therapists in 7 studies. Pharmacist interventions included medi-cation management, educational interventions, feedback to physicians, measurement of CVDrisk factors, or patient-reminder systems. Compared with usual care, pharmacist care was asso-ciated with significant reductions for systolic BP (12 studies with 1,894 patients; 26.2 mmHg[95% CI 27.8 to 24.6]); diastolic BP (9 studies with 1,496 patients; 24.5 mmHg [26.2 to22.8]); TC (8 studies with 1,280 patients;215.2 mg/dL [224.7 to25.7]); LDL cholesterol (9studies with 8,084 patients;211.7mg/dL [215.8 to27.6]); and BMI (5 studies with 751 patients;20.9 kg/m2 [21.7 to20.1]). Pharmacist care was not associated with a significant change in HDLcholesterol (6 studies with 826 patients; 0.2 mg/dL [21.9 to 2.4]).

CONCLUSIONSdThis meta-analysis supports pharmacist interventionsdalone or in col-laboration with other health care professionalsdto improve major CVD risk factors amongoutpatients with diabetes.

Diabetes Care 35:2706–2717, 2012

Cardiovascular disease (CVD) is themajor cause of death in patientswith diabetes and is about two times

more frequent in these patients comparedwith people without diabetes (1). Al-though studies have demonstrated thatcontrol of major CVD risk factors, suchas hypertension or dyslipidemia, reducesthe risk of cardiovascular complications

among patients with diabetes (2–5),CVD risk factors remain poorly con-trolled in these patients (5–7). Therefore,effective interventions and new models ofcare are needed to improve managementof CVD risk factors among patients withdiabetes, particularly in light of the in-crease in health care costs and the increas-ing problems with access to primary care

physicians in most health care systems(8). A recent systematic review evaluatingthe effect of pharmacists as teammemberson patient care in the United States hasunderlined the integration of pharmacistsas a health care team member and pro-vider of health services (9).

Pharmacists can therefore help fill thegap as primary care providers and cancontribute to the control of CVD riskfactors by their knowledge of medica-tions, their easy accessibility for patients,and their collaborative practice withphysicians (9–13). More specifically,pharmacists have the opportunity to pro-videmedication instructions to patients ateach prescription, to improve safe medi-cation use, and to assist physicians inchronic care (12,14,15). In a recentmeta-analysis, we showed that pharma-cist care improves the management ofoutpatients with major modifiable CVDrisk factors (15). However, studies con-ducted solely in outpatients with diabeteswere not included in that meta-analysis.Other reviews have evaluated pharmacistinterventions in the management of diabe-tes but focused solely on glycemic control(16), did not include only randomized con-trolled trials (RCTs), or did not pool data(17–19). Consequently, the objective ofthis systematic review and meta-analysisof RCTs is to assess the effect of pharmacistcare on major CVD risk factors among out-patients with diabetes.

RESEARCH DESIGN ANDMETHODSdWe followed the PreferredReporting Items for Systematic Reviewsand Meta-Analyses (PRISMA) statementto conduct our systematic review (20).

Inclusion criteria and outcomesWe included studies that 1) had a RCT de-sign; 2) evaluated the effect of pharmacistcare delivered by a community, hospital, orclinical pharmacist; 3) among adults out-patients with diabetes (type 1 or 2) andwith anymodifiablemajor CVD risk factors(hypertension, dyslipidemia, smoking, orobesity); and 4) compared with a usual

c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c

From the 1Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal,Quebec, Canada; the 2Institute of Social and Preventive Medicine, Lausanne University Hospital, Lausanne,Switzerland; the 3Research Institute of theMcGill University Health Centre, Montreal, Quebec, Canada; the4Institut National de Santé Publique du Québec, Montreal, Quebec, Canada; and the 5McGill Library, LifeSciences Library, McGill University, Montreal, Quebec, Canada.

Corresponding author: Valérie Santschi, [email protected] 23 February 2012 and accepted 19 June 2012.DOI: 10.2337/dc12-0369This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10

.2337/dc12-0369/-/DC1.© 2012 by the American Diabetes Association. Readers may use this article as long as the work is properly

cited, the use is educational and not for profit, and thework is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

2706 DIABETES CARE, VOLUME 35, DECEMBER 2012 care.diabetesjournals.org

M E T A - A N A L Y S I S

mailto:[email protected]

http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc12-0369/-/DC1


http://creativecommons.org/licenses/by-nc-nd/3.0/

http://creativecommons.org/licenses/by-nc-nd/3.0/

care group. Outcomeswere systolic and di-astolic BP, total cholesterol (TC), LDL cho-lesterol, HDL cholesterol, and BMI.We didnot consider fasting blood glucose or gly-cosylated hemoglobin (A1C) as outcomesbecause they are not classical CVD risk fac-tors and, in addition, were recently the sub-ject of a systematic review assessing theeffect of pharmacist care on metabolic con-trol among patients with diabetes (16).

Search strategyIn collaboration with a research medicallibrarian (A.L.C.), electronic databasesincluding MEDLINE via PubMed (1950toMarch 2012), EMBASE (1980 toMarch2012), CINAHL (1937 to March 2012),and the Cochrane Central Register of Con-trolled Trials (up to March 2012) weresearched for randomized controlled studiesof the impact of pharmacist care on majorCVD risk factors among outpatients withdiabetes, published in any language. In-clusion criteria and methods of analysiswere specified in advance and documentedin a protocol available upon request.

The PubMed search syntax served asthe basis for all search strategies, usingMedical Subject Headings (MeSH) and textterms with Boolean operators. The syntaxconsisted of three search themes inter-sected by the Boolean term “AND.”MeSHterms included cardiovascular disease-related terms (“Cardiovascular Diseases,”“Dyslipidemias,” “Diabetes Mellitus,”“Smoking,” “Overweight”) and pharmacist-related terms (“Pharmacists,” “Pharmaceu-tical Services,” “Pharmacy Service,Hospital,”“Pharmacies,” “Pharmacy”). The searchwasfocused on randomized controlled studiesusing theCochraneHighly Sensitive SearchStrategy for identifying randomized trialsin MEDLINE. The search strategy wasthen adapted for EMBASE, CINAHL, andthe CochraneCentral Register of ControlledTrials. In addition to database searches, wealso conducted a hand search of bibliog-raphies of all relevant articles. The de-tailed search strategy used is describedin the Supplementary Data online.

Methods of the reviewTwo reviewers (V.S. and A.C.) indepen-dently screened all titles and abstracts andthen examined the full text of each poten-tially eligible article using prespecified in-clusion criteria (Fig. 1).Disagreements aboutstudy selection were resolved by discussion.

Data extractionTwo reviewers (V.S. and A.C.) indepen-dently extracted data using a standardized

collection form. From each included study,the following characteristics were ab-stracted: 1) study authors and country andyear of the publication; 2) study character-istics (study setting and design, duration offollow-up, sample size); 3) characteristics ofparticipants (sex, age, CVD risk factors,medications); 4) characteristics of interven-tions (description and frequency of thepharmacist intervention); 5) characteristicsof the usual care group; and 6) outcomemeasures (change in BP, TC, LDL choles-terol, HDL cholesterol, and BMI betweenbaseline and follow-up).

Pharmacist interventions were classi-fied using a priori–defined categories basedon a recent systematic review of pharmacistcare of patients with heart failure (21) aspharmacist-directed care (pharmacist initi-ated andmanaged interventions) and phar-macist collaborative care (pharmacistcollaborated in interventions conductedby a multidisciplinary healthcare team).

Assessment of the methodologicalquality of individual studiesUsing the Cochrane Risk of Bias Tool(22), two reviewers (V.S. and A.C.) inde-pendently assessed the quality (risk ofbias) of each study regarding the follow-ing domains: adequacy of randomization,concealment of allocation, blinding of out-come assessors, completeness of data, se-lective outcome reporting, and “other bias.”

As recommended, we rated each item as 1)“low risk of bias” if it is unlikely that a biasseriously alters the results; 2) “unclear” if itis plausible that a bias raises some doubtabout the results; and 3) “high risk of bias”if it is plausible that a bias seriously weak-ens confidence in the results. Disagree-ments were resolved by discussion.

Statistical analysisStatistical analyses were conducted fol-lowing the recommendations of theCochrane Handbook for Systematic Re-views of Interventions (22) and thePRISMAstatement (20). Data were analyzed usingSTATA 11.0 software (StataCorp LP, Col-lege Station, TX). The meta-analysis wasconducted using a random-effects modelbecause of the a priori heterogeneity (15).Pooled intervention effect estimates arereported as weighted mean differenceswith 95% CI. We calculated SDs from SEsor CIs presented in the reports, if required.The x2 and I2 statistics were used to assessstatistical heterogeneity across studies, withI2 values of 50% or more indicating a sub-stantial level of heterogeneity (22). Posthoc subgroup analyses were conductedaccording to the type of pharmacist care,the type and the number of interventions,and the setting of care. Funnel plotswere drawn, and Egger tests were com-puted to explore a potential publicationbias. Sensitivity analyses were conducted

Figure 1dSelection of studies for systematic review of pharmacist care interventions. (A high-quality color representation of this figure is available in the online issue.)

care.diabetesjournals.org DIABETES CARE, VOLUME 35, DECEMBER 2012 2707

Santschi and Associates


by excluding relatively small studies andrestricting analysis to good-quality studies.

RESULTS

Results of the research and theincluded studiesInitially, 14,035 citations were identified(Fig. 1). After initial screening of titles andabstracts, 169 potentially relevant full-text

articles were reviewed for eligibility. Thereview included 15 RCTs, all published inEnglish, involving 9,111 participants.

Details of the included studies, includ-ing the characteristics of the participants,the interventions of the pharmacist, and theoutcome are given in Tables 1 and 2. The9,111 participants (52% women) were amean age of 63 years (range 49–70). Themean duration of follow-upwas 11months

(range 4–24). Participants were most oftenfollowed up in outpatient clinics (12 stud-ies); for example, Veterans Affairs MedicalCenters or primary care clinics. Three stud-ies were conducted in community pharma-cies (8,23,24). Nine of the 15 studies wereconducted in North America (7 in the U.S.and 2 in Canada), 4 in Asia (Thailand,China, United Arab Emirates, and India),1 in Europe (Spain), and 1 in Australia.

Table 1dCharacteristics of the included studies: study setting and design, sample size, and study participants

Source; country Study settingStudy design,duration

Sample size (intervention/usual care) Study participants; mean age

Pharmacist-directed careAl Mazroui et al., 2009(25); United ArabEmirates

Outpatient clinic RCT, 12 months 234 (117/117) Patients with type 2 diabetes taking oralhypoglycemic Med; 49 years

Chan et al., 2012 (26);Hong Kong

Outpatient clinic RCT, 9 months 105 (51/54) Patients with uncontrolled type 2diabetes (AIC$ 8%), taking at least 5Meds including one hypoglycemicMed; 62 years

Fornos et al., 2006(23); Spain

Communitypharmacy

RCT, 13 months 112 (56/56) Patients with diabetes taking oralantidiabetic Med; 64 years

Kraemer et al., 2012(24); U.S.

Communitypharmacy

RCT, 12 months 67 (36/31) Employees with type 1 or 2 diabetes;54 years

Phumipamorn et al.,2008 (27); Thailand

Outpatient clinic RCT, 8 months A1C, TC, HDL-C,130 (63/67); LDL-C,108 (53/55)

Patients with diabetes (A1C .7%);54 years

Planas et al., 2009(28); U.S.

Communitypharmacy

RCT, 9 months 40 (25/15) Patients with uncontrolled diabetes(A1C .7.0%) and hypertension (BP$130/80 mmHg) or takingantihypertensive Med; 65 years

Rothman et al., 2005(29); U.S.

Outpatient clinic RCT, 12 months 194 (99/95) Patients with type 2 diabetes(A1C $8.0%); 55 years

Sriram et al., 2011(30); India

Outpatient clinic RCT, 8 months 120 (60/60) Patients with type 2 diabetes; 56 years

Pharmacist-collaborativecare

Clifford et al., 2005(31); Australia

Outpatient clinic RCT, 12 months 180 (92/88) Patients with type 2 diabetes taking atleast one prescribed Med; 70 years

Edelman et al., 2010(32); U.S.

Outpatient clinic RCT, 12 months 239 (133/106) Patients with uncontrolled diabetes(A1C.7.5%) and hypertension(BP.140/90 mmHg) taking Medfor diabetes; 62 years

McLean et al., 2008(8); Canada

Communitypharmacy

RCT, 6 months 227 (115/112) Adult patients with diabetes and BP.130/80mmHg taking insulin or oralhypoglycemic Med for .6 months;65 years

Pape et al., 2011 (33);U.S.

Outpatient clinic Cluster-RCT,24 months

6,963 (2,047/4,916) Patients with type 1 or 2 diabetes;64 years

Scott et al., 2006 (34);U.S.


Simpson et al., 2011(35); Canada


Taveira et al., 2010(36); U.S.

Outpatient clinic RCT, 4 months 109 (58/51) Patients with uncontrolled type 2diabetes (A1C between 7 and 9%);64 years

HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; TC, total cholesterol; Med, medication; HTA, hypertension.


Pharmacist care in diabetes

Table

2dCha

racteristics

oftheinclud

edstud

ies:keycomponentsof

pharmacistinterventions,intervention

frequency,

descriptionof

usua

lcare

group,

andoutcom

esextracted

Source;cou

ntry

Key

compo

nentsof

pharmacistinterventio

nsIntervention

frequency

Description

ofusual

care

grou

pOutcom

esextracted

Pharmacist-directedcare

AlM

azroui

etal.,20

09(25);U

nitedArab

Emirates

Pharmaceuticalcare

including:1)

Patient

educationregardingMed

(sideeffects,storageof

medication),d

isease

(riskon

diabetes

complications),lifestyleandcompliance;distribu

tion

ofprinted

leaflet;2

)behavioralm

odificatio

nsuppo

rt(adv

iceon

self-

monitoring

ofglycem

iccontrol;ph

ysicalexercise;M

edcompliance

andsm

okingcessation);and

3)discussion

with

physicianand,

ifnecessaryrecommendation(dosagesimplification)

toph

ysician

regardingMed

changes.

Not

applicable

Usualcarefrom

medicaland

nursingstaff;no

patient

education;

notreatm

ent

plan

BP,T

C,L

DL-C,

HDL-C,and

BMI

atfollo

w-up

Chanetal.,20

12(26);

Hon

gKon

gFace-to-face

interviewwithph

armacistbefore

each

physicianvisit

including:1

)Assessm

entof

Med

history(prescriptionMed,such

asover-the-cou

nter

drugs,vitamins,andherbalsupplem

ents);2)

evaluationof

patientMed

adherence;3

)education

abou

tCVDs

andlifestylemod

ification

s;4)

distribu

tion

ofcolorstickersfor

pillb

oxes

orMed

bags

ofpatients;and5)

DRPs

identification

withno

tification

toph

ysician.

Every

patientencoun

ter

Usualcare

witho

utph

armacistintervention

Change

inBP

,HDL-C,

LDL-C,T

C,and

BMI

from

baselin

e

Fornos

etal.,20

06(23);Spain

Pharmacotherapyfollo

w-upprogram

including:1

)Patient

interviewregardingMed,h

ealthprob

lems,andlifestyle;2)

patienteducation

regardingdiabetes,lifestyle(physical

activity,h

ealthy

diet,smok

ingcessation)

andMed;3)

distribu

tion

ofverbalandwritten

inform

ationabou

tthe

correctuseof

Med;and4)

detectionandresolution

ofDRPs.

Mon

thly

Usualcare

BP,T

C,L

DL-C,

HDL-C,and

BMI

atfollo

w-up

Kraem

eretal.,20

12(24);U

.S.

Multiplescheduled

educatio

nalappo

intm

entsinclud

ing:1)

Patient

educationandem

pow

erment;and2)

progress

noteto

patient’s

physicianby

fax,e-mail,or

mailafter

each

visit.

1,2,

3mon

ths;thereafter,

every1–

3mon

ths

Usualcare

received

written

andpo

staleducational

inform

ationabou

tdiabetes

Change

inBP

,HDL-C,

LDL-C,and

TC

Phum

ipam

ornetal.,

2008

(27);T

hailand

Check

Med

pillcount;p

atient

educationregardingMed,d

iabetic

treatm

entandlifestyle;d

istributionof

diabeticpamph

let.

Every

2mon

ths(7

visits

over

12mon

ths)

Usualcare

provided

byph

ysicians;usualM

eddispensatio

nby

pharmacist

Change

inTC,L

DL-C,

andHDL-Cfrom

baselin

ePlanas

etal.,20

09(28);U

.S.

Acommun

ity-based

MTM

program

includ

ing:1)

Med

review

related

tocurrentprescribed

andno

nprescribedMed

toidentifyDRPs;

2)DRPs

identifi

catio

n;3)

ifDRPidentifi

ed,recom

mendatio

nsto

physicians

regardingadjustmentHTAMed

dose

andadditio

nMed

byfaxor

teleph

one;4)

patient

educationregardingMed,lifestyle

anddiet;and

5)acopy

ofthevisitno

tesent

toph

ysician.

Mon

thly

Usualcare

Change

inBP

andTC

from

baselin

e

Rothm

anetal.,

2005

(29);U

.S.

Intensivediabetes

managem

ent:1)

Patient

educationandcoun

seling

regardingdiabetes

andMed;2)Med

adjustments(initiatio

nor

increase)basedon

BP,cho

lesterol,and

glucoseevidence-based

algorithmswith

theapprovalof

physicians;and

3)proactive

managem

ento

fclinicalparametersusingadatabase

totrack

patient

outcom

esandproactivelyim

provecare.

Every

2–4weeks

ormore

frequentlyifnecessary

Usualcare

Change

inBP

andTC

from

baselin

e

Contin

uedon

p.2710



Tab

le2d

Con

tinu

ed

Source;cou

ntry

Key

compo

nentsof

pharmacistinterventio

nsIntervention

frequency

Description

ofusual

care

grou

pOutcom

esextracted

Sriram

etal.,20

11(30);India

Pharmaceuticalcare

including:1)

Patient

educationincluding

Med

coun

selin

g,dietary,exercise,andlifestyle

mod

ificatio

ns;2)distribu

tionof

educationalm

aterials,

including

inform

ationleafletanddiabeticdiary;and3)

contact

bytelepho

neto

mon

itorpatient

progress.

Not

applicable

Usualcare

receivingno

pharmaceuticalcare

BMIatfollo

w-up

Pharmacist-collab

orativecare

Clifford

etal.,20

05(31);A

ustralia

Pharmaceuticalcare

program

including:1)

Med

review

including

complem

entary

anddiabetes

Medsby

patientqu

estion

naire;

2)listin

gandrankingof

DRPs;3

)recommendations

toph

ysicianregardingMeds;and4)

patienteducationrelatedto

Meds,diet,exercise,andho

meglucosemon

itoring

(distributio

nof

educationalp

amph

lets).Team

mem

bers:physician

Atbaselin

e,6and12

mon

thsplus6weekly

telepho

necontacts

Usualcare

Change

inBP

,TC,

HDL-C,and

BMI

from

baselin

e

Edelm

anetal.,20

10(32);U

.S.

Ateach

GMCsession(7–8patientswith

thecare

team

):1)

Med

review

from

medicalrecordsby

pharmacistandph

ysician;2)

review

ofBP

andho

mebloo

dglucose

readings

byph

armacist

andph

ysician;3)developm

entof

individualized

plan

forMed

orlifestylemanagem

entwithph

armacistandph

ysician;4

)adjustmentof

Med

byph

armacistandph

ysicianandreportto

prim

arycare

providers;and5)

patient

educationand

coun

selin

grelatedto

Med

andlifestyle.T

eam

mem

bers:

physicianan

dnurse

Every

2mon

ths(7

visits

over

12mon

ths)

Usualcare;n

oactive

intervention

BPatfollo

w-up

McLeanetal.,20

08(8);Canada

Pharmacist-nu

rseteam

includ

ing:1)

Patienteducatio

nand

coun

selin

gregardingcardiovascularrisk

redu

ction;

2)distribu

tionof

HTAeducationpamph

letandwalletcard

documenting

recorded

patientBP

measuresfaxedto

physicians;and3)

patient’srisk

factors,currentMed

andBP

measureswith

anysuggestio

nsforfurther

managem

entbased

ongu

idelines

faxedto

physicians.Team

mem

ber:nurse

Every

6weeks

Usualcarewithdistribu

tion

ofBP

walletcard,pamph

leton

diabetes;w

ithgeneral

counselingfrom

nurseand

pharmacist

Change

inSB

Pfrom

baselin

e

Pape

etal.,20

11(33);U

.S.

Physician-ph

armacistteam

includ

ing:1)

Reviewingof

medical

chartsof

patientswithan

elevated

LDL-C;2

)recommendations

onMed

andfollo

w-uplabo

ratory

mon

itoringsent

electron

icallyto

physician;3

)if

recommendationisacceptedby

physician,telepho

nic

pharmacistintervention

topatient

(con

firm

ationof

Med

historyandprevious

adversereaction

s,identificatio

nof

barriersof

adherence);and4)

documentatio

nof

patient

communicatio

nandcare

inmedicalrecord

andconsignation

byph

ysician.T

eam

mem

ber:physician

Not

applicable

Usualcare

LDL-CandBP

atfollo

w-up

Contin

uedon

p.2711



Types of pharmacist interventionsEight studies were pharmacist-directedcare (23–30) and seven were pharmacistcollaborative care (8,31–36) (Tables 1and 2). Pharmacist interventions were 1)medication management (monitoring ofdrug therapy such as adjustment andchange of medications, medication reviewfrom patient interviews, or assessment ofmedication compliance) in 14 studies(8,23–29,31–36); 2) educational interven-tions to patients (education and counselingabout medications, lifestyle and physicalactivity or about compliance; distributionor use of educational material; educationalworkshop) in 14 studies (8,23–34,36); 3)feedback to health care professional (drug-related problems [DRPs] identification, rec-ommendation anddiscussionwithphysicianregarding medication changes or problemsof compliance, development of treatmentplans) in 12 studies (8,23–26,28,29,31–35); 4) measurement of CVD risk factors(measurement of BP or review of laboratorydata by the pharmacist during follow-up)in 3 studies (8,32,35); and 5) patient-reminder systems (telephone contact)in 5 studies (26,29,30,32,33).

Methodological quality of includedstudiesAs assessed by the Cochrane Risk of BiasTool (22), the quality of the 15 includedstudies varied (Supplementary Fig 1).Most studies adequately described random-ization sequence generation, were free ofselective outcome reporting, and were freeof other bias (e.g., important baseline im-balance in patient characteristics). Alloca-tion concealment and blinding to outcomeassessors was not described inmost studies.In all studies, participants were not blindedto pharmacist intervention.

OutcomesThe outcome was systolic BP in 12studies (including 1,894 patients)(8,23–26,28,29,31,32,34–36), diastolicBP in 9 studies (1,496 patients) (23–26,29,31,32,35,36), TC in 8 studies(1,280 patients) (23–27,29,31,35), LDLcholesterol in 9 studies (8,084 patients)(23–27,33–36), HDL cholesterol in 6 stud-ies (826 patients) (23–27,31), and BMI in 5studies (751 patients) (23,25,26,30,31).No study reported smoking as an outcome.

BPOf the 12 studies reporting systolic BP(8,23–26,28,29,31,32,34–36), 7 dem-onstrated statistically significant reduc-tions with pharmacist care. Of the nineT

able

2dCon

tinu

ed

Source;coun

try

Key

compo

nentsof

pharmacistintervention

sInterventio

nfrequency

Description

ofusual

care

grou

pOutcomes

extracted

Scottetal.,20

06(34);U

.S.

Pharmacist-managed

diabetes

care

program

(6–8patients)

includ

ing:1)

Reviewof

nutritionandbasicdiabetes

managem

ent;2)

patient

educationandcoun

selin

gabou

tlifestyle,com

pliance

anddiabetes

(goalsetting);and3)

encouragementto

mon

itor

bloo

dglucose.T

eam

mem

bers:

physician,n

urse,

anddietitian

At3,

6,and9mon

ths

Usualcare

Changein

BMIandBP

from

baselin

e

Simpson

etal.,20

11(35);C

anada

Med

review

from

patientdiscussion,

measurementof

CVDrisk

factors(BP,

height,w

eigh

tandheartrate),recommendation

sto

physicianregardingMed

managem

entof

BPandother

risk

factors,discussionandapprovalof

recommendatio

nsby

physician,implementation

ofrecommendatio

nsby

pharmacist.Team

mem

ber:p

hysician

Atph

armacists’,ph

ysicians’,

orpatients’discretio

nUsualcare;n

ocontribu

tion

sfrom

pharmacists

Changein

BP,T

C,and

LDL-Cfrom

baselin

e

Taveira

etal.,20

06(36);U

.S.

Pharmacist-ledgrou

pmedicalvisitprogram

(4to

8patients)

includ

ing:1)

Edu

cation

alintervention

(edu

cation

ofMed,

diseaseandlifestyle);and2)

pharmacologicandbehavioral

interventio

n(edu

cation

ofMed,d

isease

andlifestyle;M

edchanges;exercises,prescription

s,anddistribu

tionof

report

card

containingmedicalhistory,Med,andlabo

ratory

values).

Team

mem

bers:n

urse,nutritionist,an

dphysical

therap

ist

Once

aweek

Usualcare

Changein

BPand

LDL-Cfrom

baselin

e

GMC,groupmedicalclinicinclud

ingph

armacistand

prim

arycareinternist;HDL-C,H

DLcholesterol;HT,h

ypertension;

LDL-C,LDLcholesterol;Med,m

edication;M

TM,m

edicationtherapymanagem

ent;TC,total

cholesterol;Med,m

edication;

SBP,

systolicBP

.




studies reporting diastolic BP (23–26,29,31,32,35,36), three demonstratedstatistically significant reductions withpharmacist care. Pooled analyses indicatethat pharmacist care showed greaterreductions in systolic BP (weighted meandifference26.2mmHg [27.8 to24.6],P,0.001) and diastolic BP (weighted meandifference 24.5 mmHg [26.2 to 22.8],P, 0.001) compared with usual care (Fig.2A and B). Heterogeneity was negligiblefor systolic BP (I2 = 2%) but moderatefor diastolic BP (I2 = 46%).

TC, LDL cholesterol, and HDLcholesterolTwo of the eight studies reporting TCdemonstrated statistically significant re-ductions with pharmacist care (25,27),and the pooled estimate showed a signifi-cant reduction in TC (weighted meandifference 215.2 mg/dL [224.7 to25.7],P = 0.002; Fig. 3A). Five of the nine studiesreporting LDL cholesterol demonstratedstatistically significant reductions withpharmacist care (25–27,33,34), and thepooled estimate showed a significant reduc-tion in LDL cholesterol (weighted meandifference 211.7 mg/dL [215.8 to 27.6],P , 0.001; Fig. 3B). Six studies reportedHDL cholesterol (Fig. 3C). The pooled esti-mate did not show a statistically significantchange in HDL cholesterol (weighted meandifference 0.2 mg/dL [21.9 to 2.36], P =0.846). There was substantial heterogene-ity for TC (I2 = 75%) and HDL cholesterol(I2 = 72%) and moderate heterogeneity forLDL cholesterol (I2 = 41%).

BMIFive studies reported BMI (Fig. 3D). Twostudies (30,31) showed a statistically sig-nificant benefit of pharmacist care. Thepooled estimate showed a significant re-duction in BMI (weighted mean difference20.9 kg/m2 [21.7 to 20.1], P = 0.026)with a substantial heterogeneity betweenstudies (I2 = 92%).

Subgroup analysesWe conducted post hoc subgroup anal-yses to explore the possible differencesbetween studies according to the type ofpharmacist care, the type and the numberof interventions and the setting of care(Supplementary Table 1). These analyses,conducted only for the outcome systolic BPavailable in 12 studies, showed a statisti-cally and clinically significant reduction insystolic BP for both pharmacist-directedcare and pharmacist-collaborative care,with a tendency to a greater reduction in

systolic BP for pharmacist-directed carecompared with pharmacist-collaborativecare (weighted mean difference 28.1mmHg [212.3 to 23.7] vs. 25.7 mmHg[27.7 to23.7]). Studies conducted in com-munity pharmacies inclined to a greater re-duction in systolic BP compared with otherstudies (weighted mean difference210.0mmHg [216.4 to23.7] vs.25.5 mmHg[27.3 to23.7]). We found no major dif-ferences in systolic BP reductions accord-ing to the type and the number ofinterventions.

Publication biasWe explored the possibility of publicationbias for the 12 studies in which the out-come was systolic BP. The funnel plot forthe outcome systolic BP showed a slightasymmetry, with a borderline P value usingthe Egger test (P = 0.08), indicating a po-tential publication bias.

Sensitivity analysesBecause of a potential publication biasand to explore the impact of study qualityon the effect estimates, we conducted twosensitivity analyses. After excluding thesix smallest studies (8,25,29,31,32,35),with fewer than 80 participants per ran-domization group, the six remainingstudies demonstrated a statistically signif-icant reductions in systolic BP for phar-macist care compared with the usual caregroup (weighted mean difference 25.7mmHg [27.8 to 23.6], I2 = 0%) and ofthe same magnitude that was observedwhen all studieswere included. The secondsensitivity analysis was restricted to studiesof “good quality.” Of 12 studies with thesystolic BP outcome, 5 were of good qual-ity, with a low risk of bias on at least 4 of 6items using the Cochrane Risk of Bias Tool(22), and showed again statistically signifi-cant and similar reductions for the pharma-cist care (weighted mean difference 25.6mmHg [28.2 to22.9], I2 = 0%).

CONCLUSIONSdOur systematic re-view and meta-analysis supports the bene-fits of pharmacist interventions in themanagement of major CVD risk factorsamong outpatients with diabetes. Pharma-cist interventions were associated withsignificant reductions in systolic and di-astolic BP, TC, LDL cholesterol, and BMIcompared with the usual care group. Nosignificant change was observed with HDLcholesterol. The most frequent pharmacistinterventions were medication manage-ment (monitoring of drug therapy such asadjustment and change of medications,

medication review from patient interviews,or assessment of medication compliance),educational interventions to patients (edu-cation and counseling about medications,lifestyle and physical activity or aboutcompliance; distribution or use of educa-tionalmaterial; educationalworkshop), andfeedback to another health care professional(DRPs identification; recommendation anddiscussion with physician regarding medi-cation changes or problems of compliance;development of treatment plans).

Beyond the individual studies in-cluded in our review, no other systematicreview has examined the efficacy of phar-macist interventions in the managementof major CVD risk factors among patientswith diabetes. Nevertheless, these benefi-cial results are consistent with our recentsystematic review on the effect of pharma-cist care in CVD care among outpatients(15) and extend this earlierwork by includ-ing data from trials focusing specificallyon pharmacist care among outpatientswith diabetes.

The benefit of pharmacist interven-tions on CVD risk factors in patients withdiabetes combined with the impact ofpharmacist care on metabolic control(A1C and fasting blood glucose), as shownin a recent systematic review (16), couldtranslate in substantial benefits on patientmorbidity andmortality as well as on healthcare costs. Our review indicates that phar-macist care improves in a comprehensivemanner all major CVD risk factors amongoutpatients with diabetes, except smoking,for which no study was identified. Becausesmoking amplifies the risk of CVD amongpatients with diabetes (37) and may havedeleterious effect on diabetes control (38),studies are needed to assess the impact ofpharmacist care to reduce the risk of smok-ing among patients with diabetes. The ben-eficial effect of pharmacist care on BMI thatwe observed is encouraging because, likesmoking cessation, weight loss is difficultto achieve but of major importance amongpatients with diabetes to help reduce CVDrisk and control blood glucose (39).

We could not identify which type ofintervention was more potent to controlCVD risk factors among patients with di-abetes. Our subgroup analyses restricted tostudies assessing systolic BP did not revealclear differences in effect size according tothe type of pharmacist care, the type andthe number of interventions, or the settingof care. Moreover, we could not estimatefrom this review how much trainingthe pharmacist needs, which elementsof the intervention have a real impact, the




Figure 2dForest plots show the effect of pharmacist care on the mean difference in systolic BP (A) and in diastolic BP (B). Mean differences of lessthan 0 between pharmacist care and usual care groups indicate an effect in favor of pharmacist care. (A high-quality color representation of thisfigure is available in the online issue.)



Figure 3dForest plots show the effect of pharmacist care on the mean difference in TC (A), LDL cholesterol (B), HDL cholesterol (C), and BMI (D).Mean differences of less than 0 between pharmacist and usual care groups indicate an effect in favor of pharmacist care. (A high-quality colorrepresentation of this figure is available in the online issue.)



frequency with which the pharmacist hasto intervene, or if there is an impact onCVDrisk factors beyond the period of interven-tion. Nevertheless, identification of themore potent intervention was hamperedbecause pharmacist interventions varied

among the identified studies and oftenincluded a combination of several inter-ventions. Heterogeneity was found in theeffect of interventions on diastolic BP, TC,LDL cholesterol, and HDL cholesterol andmay be attributable to differences in terms

of interventions, care setting, and co-interventions between studies (15).

Strengths and limitationsThe strengths of our review include acomprehensive systematic review of the

Figure 3dContinued



literature, the assessment of the impact ofpharmacist interventions on all majorCVD risk factors, and the inclusion ofdifferent types of pharmacist interven-tions. Limitations include the variation inpharmacist interventions between stud-ies, preventing the definite identificationof the most beneficial intervention, andthe absence of direct evidence of the effectof pharmacist interventions on CVDevents or death among patients with di-abetes. Another limitation includes theabsence in many studies of a detaileddescription of the intervention providedto the control group. In addition, we couldnot explain the observed heterogeneity be-tween studies for some outcomes. We alsoidentified a potential publication bias.However, our sensitivity analysis restrictedto relatively large studies aswell as our anal-yses restricted to good quality studies didnot suggest any substantial bias in the esti-mation of the effect of pharmacist interven-tions. Finally,most studies were conductedin the U.S. and Canada, reflecting an ad-vanced role of the pharmacist in thesehealth care systems.

Summary and perspectivesOur review provides evidence that phar-macist interventionsdconducted aloneor in collaboration with other healthcare professionalsdimprove manage-ment of major CVD risk factors amongoutpatients with diabetes. However, ourwork also indicates that more researchis needed to assess which pharmacistinterventions are most effective, imple-mentable, and less time-consuming invarious types of health care systems orjurisdictions. Future research shouldalso demonstrate cost-effectiveness ofpharmacist interventions for patientswith diabetes and CVD risk factors. Thereis also a significant need to develop andevaluate interprofessional education ini-tiatives during pharmacy, medical, andnursing training to enhance future collab-orative care among patients with diabetes.

AcknowledgmentsdThis research was sup-ported in part by the Canadian Institutes ofHealth Research (CIHR) from the AppliedPublic Health Research Chair in Chronic Dis-ease Prevention to G.P. and by research anddevelopment funding from the Departmentof Community Medicine and Health, CHUV,Lausanne, Switzerland.The fundershadno role instudy design, data collection and analysis, deci-sion to publish, or preparation of the manuscript.No potential conflicts of interest relevant to

this article were reported.

V.S. and A.C. contributed substantially tothe conception and design of the study protocol,contributed to the acquisition and managementof data, performed statistical analysis, were re-sponsible for the analysis and interpretation ofdata, drafted the manuscript, revised the man-uscript for important intellectual content, andgave final approval for the submitted version.G.P. and B.B. contributed substantially to theconception and design of the study protocol,were responsible for the analysis and inter-pretation of data, supervised the study, obtainedfunding, revised the manuscript for importantintellectual content, and gave final approval forthe submitted version. A.L.C. contributed to theacquisition andmanagement of data, revised themanuscript for important intellectual content,and gave final approval for the submitted ver-sion. V.S. and A.C. are the guarantors of thiswork and, as such, had full access to all of thedata in the study and take responsibility for theintegrity of the data and the accuracy of the dataanalysis.

References1. Sarwar N, Gao P, Seshasai SR, et al.; Emerg-

ing Risk Factors Collaboration. Diabetesmellitus, fastingbloodglucose concentration,and risk of vascular disease: a collaborativemeta-analysis of 102 prospective studies.Lancet 2010;375:2215–2222

2. UK Prospective Diabetes Study (UKPDS)Group. Intensive blood-glucose controlwith sulphonylureas or insulin comparedwith conventional treatment and risk ofcomplications in patients with type 2 di-abetes (UKPDS 33). Lancet 1998;352:837–853

3. UK Prospective Diabetes Study Group.Tight blood pressure control and risk ofmacrovascular and microvascular com-plications in type 2 diabetes: UKPDS 38.BMJ 1998;317:703–713

4. Preis SR, Hwang SJ, Coady S, et al. Trendsin all-cause and cardiovascular diseasemortality among women and men withand without diabetes mellitus in theFramingham Heart Study, 1950 to 2005.Circulation 2009;119:1728–1735

5. Resnick HE, Foster GL, Bardsley J, RatnerRE. Achievement of American Diabetes As-sociation clinical practice recommendationsamong U.S. adults with diabetes, 1999-2002: the National Health and NutritionExamination Survey. Diabetes Care 2006;29:531–537

6. McFarlane SI, Jacober SJ, Winer N, et al.Control of cardiovascular risk factors inpatients with diabetes and hypertension aturban academic medical centers. DiabetesCare 2002;25:718–723

7. Saydah SH, Fradkin J, Cowie CC. Poorcontrol of risk factors for vascular diseaseamong adults with previously diagnoseddiabetes. JAMA 2004;291:335–342

8. McLeanDL,McAlister FA, Johnson JA, et al.;SCRIP-HTN Investigators. A randomized

trial of the effect of community pharmacistand nurse care on improving bloodpressuremanagement in patients with diabetesmellitus: study of cardiovascular risk in-tervention by pharmacists-hypertension(SCRIP-HTN). Arch Intern Med 2008;168:2355–2361

9. Chisholm-Burns MA, Kim Lee J, SpiveyCA, et al. US pharmacists’ effect as teammembers on patient care: systematic re-view and meta-analyses. Med Care 2010;48:923–933

10. Fera T, Bluml BM, Ellis WM. Diabetes TenCity Challenge: final economic and clini-cal results. J Am Pharm Assoc (2003)2009;49:383–391

11. Santschi V, Rodondi N, Bugnon O,Burnier M. Impact of electronic monitor-ing of drug adherence on blood pressurecontrol in primary care: a cluster 12-month randomised controlled study. EurJ Intern Med 2008;19:427–434

12. Tsuyuki RT, Johnson JA, Teo KK, et al. Arandomized trial of the effect of communitypharmacist intervention on cholesterol riskmanagement: the Study of CardiovascularRisk Intervention by Pharmacists (SCRIP).Arch Intern Med 2002;162:1149–1155

13. Villeneuve J, Genest J, Blais L, et al. Acluster randomized controlled Trial toEvaluate an Ambulatory primary careManagement program for patients withdyslipidemia: the TEAM study. CMAJ2010;182:447–455

14. Bogden PE, Abbott RD, Williamson P,Onopa JK, Koontz LM. Comparing stan-dard care with a physician and pharmacistteam approach for uncontrolled hyperten-sion. J Gen Intern Med 1998;13:740–745

15. SantschiV,ChioleroA,BurnandB,ColosimoAL, Paradis G. Impact of pharmacist care inthe management of cardiovascular dis-ease risk factors: a systematic review andmeta-analysis of randomized trials. ArchIntern Med 2011;171:1441–1453

16. Collins C, Limone BL, Scholle JM, ColemanCI. Effect of pharmacist intervention onglycemic control in diabetes. Diabetes ResClin Pract 2011;92:145–152

17. Evans CD, Watson E, Eurich DT, et al.Diabetes and cardiovascular disease in-terventions by community pharmacists:a systematic review. Ann Pharmacother2011;45:615–628

18. MachadoM, Bajcar J, Guzzo GC, EinarsonTR. Sensitivity of patient outcomes topharmacist interventions. Part I: system-atic review and meta-analysis in diabetesmanagement. Ann Pharmacother 2007;41:1569–1582

19. Wubben DP, Vivian EM. Effects of phar-macist outpatient interventions on adultswith diabetes mellitus: a systematic review.Pharmacotherapy 2008;28:421–436

20. Liberati A, Altman DG, Tetzlaff J, et al. ThePRISMA statement for reporting system-atic reviews and meta-analyses of studiesthat evaluate health care interventions:



explanation and elaboration. PLoS Med2009;6:e1000100

21. Koshman SL, Charrois TL, Simpson SH,McAlister FA, Tsuyuki RT. Pharmacistcare of patients with heart failure: a sys-tematic review of randomized trials. ArchIntern Med 2008;168:687–694

22. Higgins JPT, Green S. Cochrane Handbookfor Systematic Reviews of Interventions. Ver-sion 5.0.0 [updated February 2008]. TheCochrane Collaboration, 2008. Availablefrom www.cochrane-handbook.org

23. Fornos JA, Andrés NF, Andrés JC, GuerraMM, Egea B. A pharmacotherapy follow-up program in patients with type-2 di-abetes in community pharmacies in Spain.Pharm World Sci 2006;28:65–72

24. Kraemer DF, Kradjan WA, Bianco TM,Low JA. A randomized study to assessthe impact of pharmacist counseling ofemployer-based health plan beneficiarieswithdiabetes: theEMPOWERstudy. J PharmPract 2012;25:169–179

25. Al Mazroui NR, Kamal MM, Ghabash NM,Yacout TA, Kole PL, McElnay JC. Influenceof pharmaceutical care on health outcomesin patients with type 2 diabetes mellitus. BrJ Clin Pharmacol 2009;67:547–557

26. Chan CW, Siu SC, Wong CK, Lee VW. Apharmacist care program: positive impacton cardiac risk in patients with type 2diabetes. J Cardiovasc Pharmacol Ther2012;17:57–64

27. Phumipamorn S, Pongwecharak J,Soorapan S, Pattharachayakul S. Effects ofthe pharmacist’s input on glycaemic con-trol and cardiovascular risks in Muslimdiabetes. Prim Care Diabetes 2008;2:31–37

28. Planas LG, Crosby KM, Mitchell KD,Farmer KC. Evaluation of a hypertensionmedication therapy management pro-gram in patients with diabetes. J AmPharm Assoc (2003) 2009;49:164–170

29. Rothman RL, Malone R, Bryant B, et al. Arandomized trial of a primary care-baseddisease management program to improvecardiovascular risk factors and glycatedhemoglobin levels in patients with di-abetes. Am J Med 2005;118:276–284

30. Sriram S, Chack LE, Ramasamy R, GhasemiA, Ravi TK, Sabzghabaee AM. Impact ofpharmaceutical care on quality of life inpatients with type 2 diabetes mellitus. J ResMed Sci 2011;16(Suppl 1.):S412–S418

31. Clifford RM, Davis WA, Batty KT, DavisTM; Fremantle Diabetes Study. Effect of apharmaceutical care program on vascularrisk factors in type 2 diabetes: the FremantleDiabetes Study. Diabetes Care 2005;28:771–776

32. Edelman D, Fredrickson SK, Melnyk SD,et al. Medical clinics versus usual care forpatients with both diabetes and hyper-tension: a randomized trial. Ann InternMed 2010;152:689–696

33. Pape GA, Hunt JS, Butler KL, et al. Team-based care approach to cholesterol man-agement in diabetes mellitus: two-yearcluster randomized controlled trial. ArchIntern Med 2011;171:1480–1486

34. Scott DM, Boyd ST, StephanM, AugustineSC, Reardon TP. Outcomes of pharmacist-managed diabetes care services in a com-munity health center. Am J Health SystPharm 2006;63:2116–2122

35. Simpson SH, Majumdar SR, Tsuyuki RT,Lewanczuk RZ, Spooner R, Johnson JA.Effect of adding pharmacists to primarycare teams on blood pressure control inpatients with type 2 diabetes: a random-ized controlled trial. Diabetes Care 2011;34:20–26

36. Taveira TH, Friedmann PD, Cohen LB,et al. Pharmacist-led group medical ap-pointment model in type 2 diabetes. Di-abetes Educ 2010;36:109–117

37. Fagard RH. Smoking amplifies cardio-vascular risk in patients with hyperten-sion and diabetes. Diabetes Care 2009;32(Suppl. 2):S429–S431

38. Chiolero A, Faeh D, Paccaud F, Cornuz J.Consequences of smoking for bodyweight, body fat distribution, and insulinresistance. Am J Clin Nutr 2008;87:801–809

39. Fujioka K. Benefits of moderate weightloss in patients with type 2 diabetes. Di-abetes Obes Metab 2010;12:186–194



Documents

Pharmacist Interventions to Improve Cardiovascular Disease ...MedicalSubjectHeadings(MeSH)andtext terms with Boolean operators. The syntax consisted of three search themes inter-sected