This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
1School of Nursing, Midwifery and Social Work, The University of Queensland, 2Australian Centre for Evidence Based Community Care:
A Collaborating Centre of the Joanna Briggs Institute, The University of Queensland, 3School of Medicine, University of Queensland, St Lucia,
and 4Centre for Applied Nursing Research, Western Sydney University and the Ingham Applied Medical Research Institute, Liverpool, Australia
AB S T R AC T
Aim: Using themethodology of the Joanna Briggs Institute, a systematic review of current research was performed todetermine if the addition of management by nurses had been more effective in improving clinical outcomes ofpatients with type 2 diabetes attending a general practice compared with standard care.
Methods: A three-step literature search was conducted for suitable English studies with quantitative clinicaloutcomes that had been published from January 1990 to May 2014. Randomised controlled trials (RCTs) wereparticularly sought after; however, other research designs were considered. Articles were assessed by twoindependent reviewers for methodological validity, prior to inclusion in the review, using standardised criticalappraisal instruments from the Joanna Briggs Institute. When possible, quantitative data were pooled in statisticalmeta-analysis.
Results: Seven studies were of suitable quality and relevance for the review: these included three randomisedcontrol trials; two cluster- RCTs; a cluster, nonrandomised, controlled before-after study; and a cluster observationalcohort study. These studies yield evidence that nurse management in addition to standard general practitioner careleads to modest improvements in blood pressure and total cholesterol levels in adults with type 2 diabetes attendinga general practice.
Conclusion: Meta-analysis identified modest, significant improvements amongst participants in nurse manage-ment interventions (NMIs) in the following clinical outcomes: mean SBP, mean DBP and mean total cholesterol. Themajority of outcomes studied did not show any advantage to adding NMIs to general practitioner care. Two studiesreported significant improvements of participants with poor control in mean haemoglobin A1c. An RCT thatinvestigates the effect of NMIs on patients, with poor control in regard to clinical outcomes and cost effectiveness,is recommended.
Key words: diabetes mellitus type 2, general practice, meta-analysis, nurse management, practice nurse
Int J Evid Based Healthc 2016; 14:41–52.
Background
I n the year 1996, diabetes was first listed as an Aus-
tralian national health priority.1 Now 5% of Australian
city-dwellers livewith this disease, and in remoteareas, it is
estimated that one out of every 11 people has diabetes.2
Correspondence: Deborah Parker, Centre for Applied Nursing
Research, Locked bag 7103, Liverpool BC. NSW 1871, Australia.
nosed with diabetes type 2, attending a general/family
practice, were reviewed. Only studies that involved a
nurse who was both employed by the practice and
played a lead role in the intervention care were included.
Studies using a randomised controlled trial (RCT) or a
cluster RCT (cRCT) design were particularly sought after;
however, other research designs (i.e. controlled clinical
trials, interrupted time series, controlled before and after
designs and observational studies) were considered as
well. Studies were selected that reported on changes in
one ormore clinical outcomes for example, SBP, DBP, BMI,
Table 1. Results of JBI-MAStARI critical appraisal of i
JBI-MAStARI critical appraisal checklist for randomised/nonr
Reference
Blackberry et al.21 cRCT
Cleveringa et al. 200826 cRCT
Gabbay et al. 200628 RCT
Meulepas et al.27 cCBA
Taylor et al. 200320 RCT
Taylor et al. 200529 RCT
JBI-MAStARI critical appraisal checklist for comparable cohort/case-con
Reference CC
Juul et al.25
cCBA, cluster controlled before after; cOC, cluster observational cohort; cRCT, cluster Runclear if the criterion was met; Y, yes the criterion was met.
International Journal of Evidence-Based Healthcare � 2016 University
niversity of Adelaide, Joanna Briggs Institute. U
cholesterol levels, haemoglobin A1c (HbA1c) and fasting
blood glucose.
Methods of the reviewStudy appraisalA Joanna Briggs Institute (JBI) approach was used for the
review. Studies were assessed by two independent
reviewers for methodological quality using one of two
standardised critical appraisal instruments, which are
part of the JBI Meta-analysis of Statistical Assessment
and Review Instrument (JBI-MAStARI). The JBI-MAStARI
tools offer a list of criteria and reviewers assess whether
each criterion was met (yes, no or unclear). The studies
were appraised for aspects such as randomisation/allo-
cation, blinding, measure methods, group comparison
and statistical analysis. The reviewer determined if the
study had met the criterion with a yes, no or unclear. To
be included in the review, a study had to meet at least
five out of ten criteria (‘yes’ by both reviewers) for
experimental studies or nine for case control, cohort
and descriptive studies. A third reviewer was used to
resolve any disagreements. The results of this assess-
ment can be viewed in Table 1; question numbers in the
table correspond with those on the published tool.
Specific details about the interventions, study popu-
lations, methods and outcomes of significance to the
review question and specific objectives were extracted
from articles using the standardised data extraction tool
from JBI-MAStARI.
Meta-analysisWhen possible, quantitative data from the studies were
pooled in statistical meta-analysis using RevMan 5
(Cochrane Collaboration). Some studies did not report/
provide standard deviations and/or 12 month mean
data25 and were excluded from the meta-analysis. For
ncluded studies
andomised controlled trials
Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10
Y N Y N U Y Y Y Y Y
Y U U Y U N Y Y Y Y
Y U U U U Y Y Y Y Y
N U N N U Y Y Y Y Y
U U U N Y Y Y Y Y Y
Y U U Y U Y Y Y Y Y
trol studies
Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9
U Y Y Y Y Y Y Y Y
CT; N, no the criterion was not met; N/A, not applicable; Q, question; U, it was
of Adelaide, Joanna Briggs Institute 43
nauthorized reproduction of this article is prohibited.
Individual reasons for exclusion are reported in the
source review.19
The interventionsNMIs were varied and included both individual and
group education for participants on lifestyle, medication
adjustment and treatment decisions based on algor-
ithms (Table 3). In Blackberry et al.21, this education
occurred over the telephone. In the Taylor et al. study,20
practice nurses titrated medications for diabetes, cho-
lesterol and hypertension following algorithms based on
national guidelines. Two studies used a computer
decision support system.26,27 Outcomes reported in
the seven reviewed studies included one or more of
the following: HbA1c, fasting blood glucose, lipids [total
cholesterol (TC), LDL], BMI, SBP or DBP (Table 3).
Main clinical outcomesGlycated haemoglobin A1cAll seven included studies assessed HbA1c. Meta-
analysis performed on the final mean HbA1c values of
four of these studies21,26–28 did not identify any
advantage to participants who had received the NMIs
International Journal of Evidence-Based Healthcare � 2016 University
niversity of Adelaide, Joanna Briggs Institute. U
as compared with standard GP care (Fig. 2). The results of
Taylor et al.29, which were not included in the meta-
analysis because the standard deviation data were not
reported, described an advantage to participants receiv-
ing the NMI at 4 months that approached significance
(P¼ 0.10). The cluster observational cohort study per-
formed by Juul et al.25 found comparable mean HbA1c
values when comparing practices with well-imple-
mented nurse-led diabetes consultations to practices
that did not have a nurse. The study of Taylor et al.20
was excluded from the meta-analysis as it was limited to
participants with long-standing diabetes, one or more
major medical comorbid conditions and an HbA1c
greater than 10% at baseline. However, they did observe
a significant decrease in HbA1c mean in the participants
receiving NMI as compared with controls at 1 year
(P¼ 0.01, effect size 0.37).20
Several studies assessed the number of participants
who reached an HbA1c target. Meta-analysis combining
data from two such studies, one using an HbA1c target of
less than 7%26 and the other a target of less than 7.5%,27
showed that NMIs did not impact the proportion
of participants that had reached the HbA1c target at
of Adelaide, Joanna Briggs Institute 45
nauthorized reproduction of this article is prohibited.
Table 3. Included studies
Studies Method/length Participants Interventions Usual care
Blackberry et al.21
Victoria,Australia
Cluster RCT;follow-up: 18months
Sites¼ 59 GP clinics; partici-pants had diabetes type 2attending a GP; interventionn¼ 236 control n¼ 237
Practice nurses fromintervention prac-tices received 2 daysof training in a tele-phone coachingprogramme, whichaimed to deliver 8telephone and oneface-to-face coachingepisodes per patient
Usual care by their GP,which may haveincluded referral to dia-betes educators, dieti-tians and diabetessocialists as part of thestandard diabetes careof that practice
Cleveringaet al.26 theNetherlands
Cluster RCT;follow-up: 1 year
Sites¼ 55 GP clinics participantshad diabetes type 2 attend-ing a GP; interventionn¼ 1699 control n¼ 1692
A 1 hr consultationwith nurse who useda computeriseddecision support sys-tem (CDSS) withdiagnostic and treat-ment algorithms anda recall system
Usual care was providedby the GP or a nurseunder direction of theGP
Gabbay et al.28
Hershey, Penn-sylvania, USA
RCT; follow-up: 1year
Sites¼ 2 GP clinics participantshad diabetes (95% had type2) attending a GP clinic; inter-vention n¼ 150 controln¼ 182
Nurse implementedalgorithms undersupervision of GP.Also included werebehavioural goal set-ting, individualisedcare plans, self-man-agement education,follow-up phonecalls and referrals
Usual care by their GP,which might includereferrals to other prac-titioners
Sites¼ 193 GP clinics; partici-pants had diabetes type 2attending a GP; n¼ 13 117
Practices with wellimplemented nurse-led type 2 diabetesconsultations
Practices with either nonurse employed or noimplemented nurse-ledconsultations
Taylor et al.20
USARCT; follow-up: 1
yearSites¼ 1 GP clinic; participants
had long-standing diabetes(type 2¼ 97% usual care,93% intervention) with oneor more major medicalcomorbid conditions and anHbA1c greater than 10% atbaseline attending a GP;intervention n¼ 61 controln¼ 66
Individual consultationwith nurse to reviewmedical, lifestyle andpsychosocial statusand develop self-management plan.Group classes eachweek for 4 weeksusing a workbookand discussion for-mat. Telephone fol-low-up to reviewpatient goals, medi-cation use, symp-toms, glucosemonitoring, BPmonitoring and self-management activi-ties. Nurses followedalgorithms based onnational guidelinesto titrate medi-cations for diabetes,cholesterol and hy-pertension
Usual care with GP includ-ing diabetes pamphlets,Medic Alert pamphletand instructions tomaintain contact withGP and attend generaldiabetes classes at alocal medical centre.Participants told after 1year they would receivea workbook and meetwith nurse care man-ager or attend a group
D Parker et al.
46 International Journal of Evidence-Based Healthcare � 2016 University of Adelaide, Joanna Briggs Institute
Studies Method/length Participants Interventions Usual care
Taylor et al.29
CanadaRCT; follow-up: 4
monthsSites¼ 1 GP clinic participants
had type 2 diabetes attend-ing a GP; intervention n¼ 20control n¼ 19
Four or five visits bynurse at home, 1visit from a dietician(nurse present) and1 visit from exercisephysiologist. Initialvisit including assess-ment of knowledgeand education pro-vided based on thisassessment. Focus ofsubsequent meetingswas self-manage-ment. Case confer-ences with usualphysician for medi-cation adjustmentand advice
Usual care by GP,included one quarterlyscheduled visit
Meulepas et al.27
the NetherlandsCluster controlled
before-afterstudy/1 year
GP clinics¼ 51; patients withtype 2 diabetes attending aprimary care clinic; interven-tion n¼ 431 control n¼ 469
Nurse provided infor-mation on lifestylemanagement for riskfactors and adjustedmedication at quar-terly visits. Also usedthe diabetes supportservice (DSS), apatient recall regis-ter. The before meandata were compar-able for HbA1c(intervention, 7.3; SD1.2 and control, 7.2;SD 1.1); for SBP(intervention, 149;SD 22.9 and control,150; SD 23.3); forDBP (intervention,84;SD 12.1 and con-trol, 84; SD 12.5); fortotal cholesterol(intervention, 5.0; SD1.0 and control, 5.1;SD 0.9). The inter-vention and controlgroups each had82% smokers, amean BMI of 29 kg/m2, a mean HDLratio of 4.3 and com-parable mean trigly-cerides (intervention,1.8; SD 1.2 and con-trol, 1.9; SD 1.2)
Care by GP and DSS
cCBA, cluster controlled before after; cOC, cluster observational cohort; cRCT, cluster RCT; GP, general practice; HbA1c, glycated haemoglobin.
Figure 2. When compared to usual care, nurse-led interventions did not improve HbA1c but were more effective at 1 year inreducing mean SBP, DBP and total cholesterol.
D Parker et al.
48 International Journal of Evidence-Based Healthcare � 2016 University of Adelaide, Joanna Briggs Institute