promoting access to White Rose research papers
White Rose Research Online
Universities of Leeds, Sheffield and York
This is the author’s version of an article published in International Journal of
White Rose Research Online URL for this paper:
Ndosi, M, Lewis, M, Hale, C, Quinn, H, Ryan, S, Emery, P, Bird, H and Hill, J
(2011) A randomised, controlled study of outcome and cost effectiveness for RA
patients attending nurse-led rheumatology clinics: Study protocol of an ongoing
nationwide multi-centre study. International Journal of Nursing Studies, 48 (8).
995 - 1001. ISSN 0020-7489
The effectiveness of nurse-led care in people with rheumatoid arthritis: A systematic review
1. Mwidimi Ndosi, MSc, BSc (Hons), RN
2. Karen Vinall, MSc, MSc, BSc (Hons), MBPsS
3. Prof. Claire Hale, PhD, BA, RN
4. Prof Howard Bird, MA, MD, FRCP
5. Dr. Jackie Hill, PhD, MPhil, RN, FRCN
Academic & Clinical Unit for Musculoskeletal Nursing
Section of Musculoskeletal Disease
2ndFloor, Chapel Allerton Hospital
Leeds, LS7 4SA, UK
Effects of Nurse-led care in RA: A systematic review.
0.1 Background: The rise in the number of patients with arthritis coupled with understaffing
of medical services has seen the deployment of Clinical Nurse Specialists in running
nurse-led clinics (NLC) alongside the rheumatologist clinics. There are no systematic
reviews of NLC effectiveness in rheumatoid arthritis. Few published RCTs exist and they
have shown positive results for NLC but they have several limitations and there has been
no economic assessment of rheumatology NLC in the UK.
0.2 Objective: This paper outlines the study protocol and methodology currently being used
to evaluate the outcomes and cost effectiveness for patients attending rheumatology
0.3 Design and methods: A multi-centred, pragmatic randomised controlled trial with a non-
inferiority design; the null hypothesis being that of ‘inferiority’ of nurse-led clinics
compared to physician-led clinics. The primary outcome is rheumatoid arthritis disease
activity (measured by DAS28 score) and secondary outcomes are quality of life, self-
efficacy, disability, psychological well-being, satisfaction, pain, fatigue and stiffness. Cost
effectiveness will be measured using the EQ-5D, DAS28 and cost profile for each centre.
0.4 Power calculations: In this trial, a DAS28 change of 0.6 is considered to be the threshold
for clinical distinction of ‘inferiority’. A sample size of 180 participants (90 per treatment
arm) is needed to reject the null hypothesis of ‘inferiority’, given 90% power. Primary
analysis will focus on 2-sided 95% confidence interval evaluation of between-group
differences in DAS28 change scores averaged over 4 equidistant follow up time points
(13, 26, 39 and 52 weeks). Cost effectiveness will be evaluated assessing the joint
parameterisation of costs and effects.
0.5 Results: The study started in July 2007 and the results are expected after July 2011.
0.6 Trial registration: The International Standard Randomised Controlled Trial Number
What is already known about this topic?
In the UK, Rheumatology Clinical Nurse Specialists conduct nurse-led clinics which provide
follow-up care to patients with rheumatoid arthritis including monitoring, patient education
and psychosocial support.
Research has shown positive outcomes of nurse-led clinics but the UK studies have several
limitations and lack generalisability and evidence of cost-effectiveness.
What this paper adds?
This paper outlines a protocol and methodology of a randomised controlled trial being
conducted to demonstrate the outcomes and cost-effectiveness of nurse-led care in
This is the first UK multi-centred RCT of effectiveness of nurse-led care in rheumatology.
This is the first UK study of cost-effectiveness of nurse-led care in rheumatology.
Nurse-led care; Clinical Nurse Specialists; Rheumatoid arthritis; Randomised Controlled Trial;
Economic evaluation; Protocol.
During the past 20 years the rise in the number of patients with arthritis, understaffing of
medical services and reduced junior hospital doctors’ working hours have prompted the
rheumatology community in the United Kingdom (UK) to reassess how patient services are
provided. These pressures have meant that whilst essential medical provision remains
intact, it is often at the expense of the psychological, social, rehabilitative and educational
needs that are so necessary to enhance patient outcomes (Mounce and Ryan, 2001). To
counter these problems, rheumatology units increasingly augment the multidisciplinary
team with clinical nurse specialists who are senior nurses, specially trained to undertake
extended roles. By taking on some of the technical and patient management activities that
were previously the sole responsibility of rheumatologists, these nurses allow
rheumatologists to concentrate on the more complex tasks such as differential diagnosis
for which they are uniquely trained (Bird, 1983).
Despite this innovative development, the evidence of effectiveness of nurse-led
rheumatology clinics is limited. Our literature search produced no systematic review of
nurse-led care effectiveness in rheumatology. Yet there were several systematic reviews of
effectiveness in other chronic diseases such as diabetes (Carey and Courtenay, 2007),
coronary heart diseases (Page et al., 2005) and chronic obstructive pulmonary disease
(Sridhar et al., 2008). Our systematic review (Ndosi et al., 2010) revealed seven randomised
controlled trials (RCTs) of effectiveness of nurse-led care in rheumatology; four of which
were in rheumatoid arthritis (Hill et al., 1994, Hill et al., 2003, Ryan et al., 2006, Tijhuis et
al., 2003), two in osteoarthritis (Hill et al., 2009, Victor et al., 2005) and one in fibromyalgia
(Kroese et al., 2008). In addition there was an economic evaluation of nurse-led care in
rheumatoid arthritis (Van Der Hout et al., 2003). Of the seven RCTs of effectiveness, two
were from the Netherlands and five were from the UK. The Dutch team also undertook the
The majority of the RCTs were in rheumatoid arthritis as this is the disease that the
majority of nurses are involved with. The outcomes of patients under nurse-led care were
compared to those of the rheumatologist (Hill et al., 1994, Hill et al., 2003), in-patient team
and day-patient team care (Tijhuis et al., 2003) and staff nurse working under a
rheumatologist (Ryan et al., 2006). The primary outcomes in these studies were disease
activity, functional status, health status and coping with rheumatoid arthritis. The results of
disease activity showed that the effects of nurse-led care were not significantly different
from those of the comparators (Hill et al., 1994, Hill et al., 2003, Tijhuis et al., 2003) and in
one study nurse-led care had better effects on disease activity (Ryan et al., 2006). The
effects of nurse-led care on functional status were not different from those of in-patient
team or day-patient team care (Tijhuis et al., 2003). Nurse-led care demonstrated better
effects than standard care on health status and coping with rheumatoid arthritis (Ryan et
al., 2006). The economic analysis (Van Der Hout et al., 2003) concluded that nurse-led care
provided equivalent quality of life and utility at a lower cost.
The osteoarthritis RCTs, demonstrated better nurse-led care effects in pain control (Hill et
al., 2009) and no difference in coping with arthritis (Victor et al., 2005). In diagnosing
fibromyalgia, nurse-led diagnosis showed excellent agreement with that of the
rheumatologist and this agreement was maintained over 24 months (Kroese et al., 2008).
Despite showing positive results, the RCTs of effectiveness of nurse-led care in rheumatoid
arthritis have several limitations. The Dutch study (Tijhuis et al., 2003) was the only
multicentre study but it did not compare like with like as there was a disparity in favour of
day care and in-patient treatments in the number of visits, hours of treatment and
intensity of care. Also, the nurse-led care cohort was less impaired and had a better quality
of life at the start of the study, making it more difficult to demonstrate a significant
difference between the groups on completion. The sequential studies undertaken in Leeds
(Hill et al., 1994, Hill et al., 2009, Hill et al., 2003) and Stoke-On-Trent (Ryan et al., 2006) are
the only work to date that validates nurse-led rheumatology clinics in the UK.
Unfortunately, these studies also have their limitations, as all were undertaken in one clinic
managed by the same rheumatology nurse, one consultant rheumatologist and four junior
doctors. In addition, the sample sizes were small and there is no confirmation that these
results are reproducible from other nurse-led clinics in the UK. Finally, there has been no
economic assessment of nurse-led rheumatology clinics in the UK and so we do not know if
they are economically effective.
The aims of this study are to establish whether nurse-led rheumatology clinics are clinically
effective and cost effective for patients with rheumatoid arthritis.
2.1 Study design
This is a pragmatic trial in a complex clinical environment and is conducted over 4 years as a
multi-centred RCT. Patients are randomised using an office hours remote secure telephone
randomisation service provided by the Clinical Trials and Research Unit, University of Leeds.
Patients who fulfil the eligibility criteria and have provided written informed consent are
randomised on a 1:1 basis to either Nurse-led clinic (Experimental group) or a
rheumatologist-led clinic (Control Group). Randomisation is by random permuted blocks,
using the stratification factors, centre and DAS28 score (low ≤3.2 or moderate to severe
>3.2) at baseline. After recruitment, patients have 5 follow-up visits over 12 months. (Figure
1). The methods are consistent with current guidelines on design, conduct and analysis of
pragmatic randomised clinical trials (Moher et al., 2010, Ramsey et al., 2005, Schulz et al.,
2010, Zwarenstein et al., 2008), and those explicitly for non-inferiority trials (Bosmans et al.,
2008, Piaggio et al., 2006).
2.2 Study population:
The study comprises 180 patients with rheumatoid arthritis recruited from 10 rheumatology
centres throughout the UK. Patients with both stable and active disease are included as this
reflects the practice of clinical nurse specialists in the UK and Europe and so make the results
more meaningful. All patients are seen in one of the 10 participating centres. The centres
are distributed throughout the UK and this provides a socio-demographic mix of patients.
Inclusion criteria are: a positive diagnosis of rheumatoid arthritis as defined by the American
Rheumatism Association (Arnett et al., 1988); aged 18 years or above, and ability to
complete questionnaires unaided. Exclusion criteria are: patients unwilling to be randomised
to a nurse-led clinic or rheumatologist-led clinic; patients suffering from unstabilised
concomitant disease; patients awaiting surgery and patients who have already received care
from the practitioners involved in the study.
The hypothesis is that the outcomes from nurse-led clinics will not be inferior to those
obtained by the rheumatologist-led clinics, but at a lower cost and greater patient
Following randomisation, patients are given appointments with their respective practitioners
for weeks 0, 13, 26, 39 and 52. When patients arrive at the clinic, they are seen by an
independent assessor who oversees the completion of pain Visual Analogue Scale (VAS),
fatigue VAS, the length of morning stiffness and performs joint counts for Disease Activity
Score (DAS28). The independent assessor also gives the patient their blood form,
questionnaires in a freepost return envelope and sends them to the waiting area ready to
see their allocated practitioner. The joint examination for DAS28 can be prone to inter-
observer variation and training or agreement sessions have been shown to minimise this
(Grunke et al., 2010, Scott et al., 1996). Therefore a training session was conducted (by MN)
with the independent assessors at the study set-up meeting to ensure standardisation. The
joint examination technique based on the European League Against Rheumatism handbook
of clinical assessment in rheumatoid arthritis (van Riel and Scott, 2000). Since this is a
pragmatic trial, the practitioners in both arms of the trial (Clinical Nurse Specialists and the
rheumatologists) did not receive any more training; they manage their patients according to
their normal practice.
During the consultation, the clinical nurse specialists record their interventions in a standard
“consultation checklist” especially designed for this study. The nurse-led care interventions
may include: pain control, medication and dosage changes, intra-articular or intra-muscular
steroid injections, provision of patient education or psychosocial support, prescription of
splints, non-protocol blood tests or radiographic examination. Other interventions such as
referral to the admission ward, to the rheumatologist, physiotherapists, podiatrist or any
other health care professional may be carried out as appropriate. The referrals, conferrals
and the length of consultation are also recorded.
Patients randomised to the rheumatologist-led care (control group) also have the same
number of study visits and the rheumatologist provides care as per normal practice noting all
the interventions and referrals in the consultation checklist. The patient is then given an
appointment for the next visit.
4. Outcome measures
4.1. Primary and secondary outcomes
The primary outcome measure is the DAS28 (Prevoo et al., 1995), an internationally
recognised measurement of disease activity in rheumatoid arthritis. It is a composite
measurementcomprising objective(numberof swollenjoints and erythrocyte
sedimentation rate [ESR] or C-Reactive Protein [CRP]) and subjective (number of tender
joints and patients global assessment) indices. DAS28 score has been shown to be a strong
predictor of physical disability and radiological progression and a sensitive discriminator
between patients with high and low disease activity (Prevoo et al., 1996, van Gestel et al.,
1998). DAS28 score is widely used in making decisions about treatment effectiveness and it
forms the basis for the European League Against Rheumatism response criteria for
rheumatoid arthritis (Van Gestel et al., 1996). Scores can range from 0-9.4, assuming that an
ESR of 100 is taken as the upper limit. Levels of disease activity are defined as DAS28 ≤3.2
mild; DAS28 >3.2 and ≤5.1 moderate; DAS28 >5.1 severe (van Gestel et al., 1998, Van Gestel
et al., 1996). The DAS28 will be measured at baseline and 13, 26, 39 and 52 weeks.
Secondary measures include haematological, clinical and questionnaire data. The
haematological and clinical measures comprise: CRP or ESR, pain intensity (using 10cm VAS),
length of morning stiffness (hours/minutes) and Fatigue (10cm VAS). The following
questionnaires will be administered at baseline, 26 and 52 weeks:
Health Assessment Questionnaire (Kirwan and Reeback, 1986)
Hospital Anxiety and Depression Scale (Zigmon and Snaith, 1983)
Leeds Satisfaction Questionnaire (Hill et al., 1992)
The Arthritis Self Efficacy Scale (Lorig et al., 1989)
Rheumatoid Arthritis Quality of life Questionnaire (De Jong et al., 1997)
An additional questionnaire, the EQ-5D (The Euroqol group, 1990) will be completed at
baseline and weeks 13, 26, 39 and 52 to provide health data for the economic analysis
4.2. Power calculations
A change in DAS28 score of 1.2 or more is deemed a clinically significant improvement, and a
change of 0.6 or more reflects a moderate improvement (van Riel et al., 1996). A change in
DAS28 score of 0.6 is assigned as the ‘inferiority’/’non-inferiority’ margin. Thus:
Null hypothesis (inferiority): mean ∆DAS28RLC – mean ∆DAS28NLC ≥ 0.6.
Alternative hypothesis (non-inferiority): mean ∆DAS28RLC – mean ∆DAS28NLC< 0.6.
Where ∆ = change in, RLC = Rheumatologist-led clinic and NLC = Nurse-led clinic.
A total sample size of 180 participants (90 per treatment arm) is needed to reject the null
hypothesis of ‘inferiority’, given 90% power and 1-sided statistical testing with 2.5%
significance level (and a pre-hypothesized standard deviation in DAS28 change scores of 1.5).
This total sample size assumes a 10% drop out/non-response rate. The calculation is based
on a repeated-measures (pooled) analysis of between-group differences averaged over 4
equidistant follow up time points (13, 26, 39 and 52 weeks). The sample size calculation
assumes that the intra-class correlation coefficient (for the correlation of observations over
time within individuals) will be about 0.5.
5. Statistical analysis
5.1. Statistical analysis
Analyses will be carried out using both intention to treat and per protocol methods as
advocated in extended CONSORT guidelines (Piaggio et al., 2006). Difference in mean
summary scores will be presented with a 95% two-sided confidence interval from which we
can draw a conclusion as to whether to accept or reject the null hypothesis of ‘inferiority’
regarding the nurse-led care intervention compared to the rheumatologist-led care. The
primary outcome measure (DAS28 change score) and secondary measures will be compared
between the two groups using data pooled over time; the evaluation focusing on the
comparability of average change in DAS28 over the assessed follow up period. Secondary
measurement will focus on individual time points. Analysis will be by hierarchical repeated
measures modelling. Analysis will adjust for age, gender, centre, baseline DAS28 (and
corresponding baseline values for secondary outcomes). Multiple imputation will be used to
address the issue of missing data (Schafer, 1999). Analysis will also be carried out
investigating outcome in relation to the interaction of intervention group and specific
baseline variables: age, gender and DAS28.
5.2. Economic evaluation
The economic assessment will encompass both a cost utility analysis and a cost effectiveness
analysis (Bosmans et al., 2008, Ramsey et al., 2005). A tiered approach to the evaluation will
encompass the following economic perspectives: (i) NHS; (ii) healthcare [NHS plus direct
patient costs]; (iii) societal [direct (healthcare) plus indirect (productivity) costs].
Healthcare resource use, specifically in relation to rheumatoid arthritis, is derived through
clinic audits and follow up patient questionnaires, and embraces health professional
consultations (primary and secondary care), hospital admissions (day care, inpatient stays,
A&E visits), investigations, and treatments including over-the-counter medications. Costs will
be derived from sources of ‘national average’ costs (Curtis, 2009, NHS Executive, 2009), and
also by direct elicitation from the self-report questionnaires for private out-of-pocket
expenditure on health care service use, travel, medication, aids and special dietary
requirements. Data collected from each patient on employment status and job title
(classified according to its socioeconomic classification using the Office of National Statistics
(ONS) approach (ONS, 2000, ONS, 2000)) will be used to determine productivity losses using
the human capital approach by multiplying a patient’s reported number of days off work by
the expected average daily wage extracted from National Statistics survey databases (ONS,
2004). Multiple imputation will be used to address missing cost data (Schafer, 1999). Health
outcomes will be assessed through QALYs derived from the EQ-5D for the cost utility
analysis, and the DAS28 change score for the cost effectiveness analysis (with incremental
differences in cost being evaluated in relation to the non-inferiority margin of 0.6 in mean
Between-group economic comparisons will focus on the joint estimation of incremental costs
and effects. The precision of the estimates will be ascertained by calculating confidence
intervals around effect and cost differences, derived through adjusted linear regression
modelling. Data for costs are usually right-skewed, and will therefore be analysed using the
preferred bootstrap technique (Heyse et al., 2001, Lambert et al., 1998, Mullner, 2003).
Uncertainty around the cost-effect estimates will be shown graphically using cost
effectiveness/utility planes (Briggs and Fenn, 1998). Cost-acceptability curves will be used to
help make informative decisions regarding the cost-effectiveness of the nurse clinics at
variable ceiling willingness-to-pay cost thresholds (Stinnett and Mullahy, 1998) Sensitivity
analyses will establish the robustness of findings to various assumptions (e.g. imputed versus
5.3. Study registration and ethical approval
The study is registered as a clinical trial at the International Standard Randomised Controlled
Trial Number Register (ISRCTN29803766). Multicentre ethical approval was obtained from
the Leeds West Research Ethics Committee and site specific approvals were obtained from
Local Research Ethics Committees of the 10 participating centres. The study is being
conducted in accordance with good clinical practice in research to the Research Governance
Framework for Health and Social Care (Department of Health, 2005).
This research is carried out by funding from the Arthritis Research UK grant. It is hosted by
the University of Leeds and it has been adopted by the National Institute for Health Research
Clinical Research Network.
5.5. Conflict of interest
The authors declare that they have no conflict of interest.
6. References Download full-text
Arnett, F., Edworthy, S., Bloch, D., 1988. The American Rheumatism Association - 1987. Revised
Criteria for the classification of rheumatoid arthritis. Arthritis and Rheumatism 31, 315-324.
Bird, H.A., 1983. Divided rheumatological care: the advent of the nurse practitioner? Annals of
Rheumatic Diseases 42 (3), 354-355.
Bosmans, J.E., De Bruijne, M.C., Van Hout, H.P.J., Hermens, M.L.M., Adèr, H.J., Van Tulder, M.W.,
2008. Practical Guidelines for Economic Evaluations Alongside Equivalence Trials. Value in
Health 11 (2), 251-258.
Briggs, A., Fenn, P., 1998. Confidence intervals or surfaces? Uncertainty on the cost-effectiveness
plane. Health Economics 7 (8), 723-740.
Carey, N., Courtenay, M., 2007. A review of the activity and effects of nurse-led care in diabetes.
Journal of Clinical Nursing 16 (11c), 296-304.
Curtis, L., 2009. Unit costs of health and social care. Canterbury: Personal Social Services Research
Unit, University of Kent.
De Jong, Z., Van Der Heijde, D., McKenna, S., 1997. The reliability and construct validity of the
RAQoL: a rheumatoid arthritis-specific quality of life instrument. British Journal of
Rheumatology 36, 878-883.
Department of Health, 2005. Research Governance Framework for Health and Social Care. Crown
Grunke, M., Antoni, C., Kavanaugh, A., Hildebrand, V., Dechant, C., Schett, G., Manger, B.,
Ronneberger, M., 2010. Standardization of Joint Examination Technique Leads to a
Significant Decrease in Variability Among Different Examiners. The Journal of Rheumatology
37 (4), 860.
Heyse, J., Cook, J., Carides, G., 2001. Statistical considerations in analysing healthcare resource
utilization and cost data. In: Drummond M, McGuire, A. (Eds.), Economic Evaluation in
Healthcare Merging Theory with Practice. Oxford University Press, Oxford.
Hill, J., Bird, H., Harmer, R., 1994. An evaluation of the effectiveness, safety and acceptability of a
nurse practitioner in a rheumatology outpatient clinic. British Journal of Rheumatology 33,