healthmedicinet

The methods of this study are described according to the CONSORT statement for cluster-randomised
trials 19].

Study design

This study is a stepped-wedge cluster-randomised controlled trial. A multifaceted
implementation strategy for the Dutch multidisciplinary care guideline for nonspecific
LBP will be compared to passive dissemination of the guideline. The implementation
strategy will be targeted at both health care professionals (i.e., GPs, OPs and PTs)
and patients. The HCPs are allocated to one of four clusters, which are compiled based
on the HCPs’ geographic proximity to each other. This grouping allows for minimisation
of contamination between the participating HCPs. A stepped-wedge design is applied
to assure stepwise implementation of the guideline in all of the participating HCPs
(Fig. 1). Patients are allocated according to their GP/PT allocation, i.e., patients registered
within a practice that is in the control group at time of enrolment will automatically
be allocated into the control group for patients.

Fig. 1. Design of the stepped-wedge trial

The Medical Ethics Committee of the VU University medical centre assessed this study
design and procedures, and in accordance with the local regulatory guidelines and
standards for human subjects protection in the Netherlands (Medical Research Involving
Human Subjects Act [WMO], 2005), this study proved to be exempt from further medical
ethical review.

Participants

Health care providers

This study comprises 25 GP practices (accounting for 53 individual GPs), 19 PT practices
(accounting for 43 individual PTs), and 37 OPs. Inclusion criteria for health care
providers are: practising within the municipality of Amsterdam, and regularly working
with patients with LBP. Participating HCPs are allocated to one of four clusters based
on defined geographical areas, which is the eligibility criterion for clusters in
this trial.

Patients

Five hundred patients with LBP will be included to participate in this study. All
patients registered in the practices of participating GPs and PTs receive an information
leaflet about this study, and a recruitment letter from their GP/PT. Furthermore;
PTs have information and recruitment posters and leaflets in their practices. Patients
interested in participation can apply directly via e-mail or using a postal reply
card at the research assistants’ office, where eligibility criteria are assessed prior
to inclusion.

Patients are eligible if they have a minimum age of 18 and a maximum age of 75 years,
have access to the Internet, have visited their GP or PT due to back complaints no
later than 3 months prior to inclusion, and are diagnosed with nonspecific LBP. Nonspecific
LBP is defined as low back pain (with or without motor and/or sensory deficits in
one or both legs) that is not caused by underlying specific pathology (red flags),
i.e., a tumour, (osteoporotic) vertebral fracture, ankylosing spondylitis, and cauda
equina syndrome.

Patients having the following characteristics will be excluded from this study: serious
long existing comorbidity, i.e., Alzheimer’s disease, Multiple Sclerosis, Parkinson’s
disease, ALS, CVA (diagnosed 1 year prior to inclusion up to moment of inclusion),
confirmed pregnancy (identified 1 year prior to inclusion up to moment of inclusion),
malignancy (diagnosed 5 years prior to inclusion up to moment of inclusion), and severe
psychiatric disorders, i.e., schizophrenia and bipolar disorder.

In accordance with the local regulatory guidelines and standards for human subjects
protection in the Netherlands (Medical Research Involving Human Subjects Act [WMO],
2005), and the assessment of the Medical Ethics Committee of the VU University medical
centre, no written informed consent for participation was necessary to be obtained
from participants.

Interventions

Passive dissemination of the guideline

In this study, a multifaceted implementation strategy (intervention group) targeted
at both the patients and the professionals will be compared to passive dissemination
of the guideline (control group). For HCPs, passive dissemination encompasses digital
dissemination of the guideline (i.e., HCPs will receive an e-mail with this guideline
in PDF format attached). Patients are offered a website on which solely the brief
patient information from the guideline is published.

Multifaceted implementation strategy for health care providers

HCP clusters in the intervention arm will be invited to participate in a multidisciplinary
continuing medical education (CME) training session that is developed in close collaboration
with a professional educationalist. The training module meets the educational requirements
of the Dutch physicians and physiotherapists associations.

The training module focuses on improving collaboration and communication between various
HCPs, which is practiced by means of a so called barriers and facilitators carousel,
in which several small groups of HCPs first discuss the barriers they encounter in
collaborating and communicating with each other. The groups then interchange these
barriers and discuss strategies to overcome them. In a plenary session, the HCPs then
exchange their group findings and agree upon a set of practical strategies for dealing
with the encountered barriers. Furthermore, communication skills needed for managing
patients with nonspecific LBP are trained by means of a case study that dictates a
role-play. The case study presents a complicated case of LBP in which yellow flags
(i.e., psychosocial risk factors) and blue flags (i.e., occupational risk factors)
play an important role. In this role-play, the HCPs swap professions with each other
and play a role from the perspective of another HCP, allowing them to think outside
their own frames of reference, while they also learn to identify important yellow
and blue flags and act upon them in a multidisciplinary manner. During the role-play,
the HCPs are encouraged to practice the previously agreed upon strategies. After the
training module, HCPs repeatedly (directly after the training, and after 3, 6, and
12 months) receive a reminder about the strategies by e-mail, along with a social
map of contact details of all HCPs that attended the specific training. The training
is conducted one time per cluster and takes 2,5 h. It is organized and given by at
least one member of the research team (in order to assure scientific quality), and
one practicing HCP (to ensure relevance and connection to daily practice).

Finally, HCPs will gain access to an interactive website, containing information on
LBP and LBP guidelines, updates on the study, and a HCP forum.

Multifaceted implementation strategy for patients

Patients randomised into the intervention arm will gain access to an informative website
aimed at reducing patients’ negative back beliefs and improve their cognitions. The
website provides comprehensive information about LBP, such as practical advices (e.g.,
on self-management), working and returning to work with LBP, exercise tips and possibilities
to contact researchers, HCPs, and other patients by means of a forum and social media.
An important part of the implementation strategy and this website are short video
messages in which actors and HCPs share their fictional experience with LBP and provide
tips on self-management of and working with LBP. These videos are based on the effective
Australian mass media campaign ‘Back Pain: Don’t Take It Lying Down’ 20]. The website is also available in a mobile version to be visited on any electronic
device, e.g., a smartphone or tablet.

Objectives

The primary objective of this study is to evaluate the (cost-) effectiveness of the
multifaceted, patient and professional based strategy for implementation of the Dutch
multidisciplinary care guideline for nonspecific LBP on patients’ back pain beliefs
and expectations. The secondary objective is the improvement of functional status
and reduction of (work) disability of patients with LBP, and improvement of guideline
adherence by HCPs.

Outcomes

Outcomes on patient level

The primary outcome measure of this study is the back beliefs and expectations of
patients with LBP, which will be assessed on individual level using the Back Beliefs
Questionnaire 21]. Secondary outcome measures on patient level include functional status measured with
the RDQ-24; quality of life measured with the EuroQol questionnaire; level of pain
measured using an adapted form of the PCI questionnaire, and health care utility and
productivity losses measured using the PRODISQ and TIC-P questionnaires. Measurements
at patient level will take place at baseline and after 3, 6, and 12 months follow-up.

Outcomes on HCP level

Health care provider outcomes are measured at the level of the individual HCP. Using
questionnaires, HCP knowledge of and attitudes toward the guideline will be measured,
as well as level of perceived self-efficacy on multidisciplinary communication and
collaboration. Measurements on HCP level will take place at baseline and after 3,
6 and 12 months follow-up.

Furthermore, guideline adherence among GPs will be assessed using performance indicators
(Table 1). These professional behaviours are considered indicators for adherence to this guideline,
because the guideline recommends a watchful waiting approach for acute LBP, and a
multidisciplinary approach to treatment for chronic LBP, while it discourages (early)
referrals of LBP patients for diagnostics and specialist medical care.

Table 1. Performance indicators to measure guideline adherence among GPs

To enhance quality of the data, all data from questionnaires will be collected using
online questionnaires, which are programmed to reduce impossible and missing values.
Data on performance indicators will be gathered using software developed to select
all GP reports on consultations with patients having LBP.

Process evaluation

The Linnan and Steckler framework for process evaluation of public health interventions
and research will be used to evaluate the implementation process for this study at
patient and HCP level 22]. The feasibility of the implementation strategy, barriers and facilitators for implementing
the guideline, and the satisfaction of the participants (HCPs as well as patients)
with the intervention are measured using questionnaires. In order to gain more in-depth
knowledge on the satisfaction and experiences of the participants, semi-structured
qualitative interviews with both HCPs and patients are conducted.

Subgroup analysis

Subsequent to the process evaluation, a subgroup evaluation analysis amongst participating
patients will be performed. Using an explorative qualitative design with semi-structured
interviews, 10 patients from four ethnic backgrounds (i.e., Dutch, Moroccan, Turkish,
and Surinam, which are the most common ethnicities in the city of Amsterdam) will
be interviewed to gain insight into their experiences with this study (e.g., the recruitment
process, the intervention received). The differences in experiences between these
ethnic groups will be mapped in order to gain insight into possible barriers and facilitators
in involving various ethnic groups in future (implementation) research.

Sample size

The sample size calculation is based on a hypothesized 10 % improvement of the primary
outcome measure back beliefs of LBP patients (measured using the Back Beliefs Questionnaire).
An intra-class correlation coefficient (ICC) of 0.05 is applied to adjust for the
cluster randomisation design. Assuming a 10 % improvement of a mean score of 26.5
(95%BI 26.1-26.8, SD 6) on the Back Beliefs Questionnaire, and applying an ICC of
0.05, the necessary sample size amounts to 500 patients. This calculation takes into
account a dropout-rate of 20 %, power (1-beta) of 0.90 and an alpha of 0.05. The sample
size is calculated at patient level, because the primary outcome measure of this study
is back beliefs of patients.

Randomisation

HCPs will be assigned to one of the four clusters. The clusters sequentially receive
the intervention, and the moment of intervention rollout within one cluster the unit
of randomisation is (Fig. 1). Randomisation is performed by means of computer-generated allocation. Patients
are blinded to group assignment.

The allocation is not concealed from HCPs or for the researchers, because the start
of the intervention (i.e., organising or receiving the CME training) is an obvious
point in time that does not lend itself for blinding. An independent research assistant
will perform the allocation sequence, enrolling of participants, and assignment of
participants to groups.

Statistical methods

Data on outcome measures will be gathered at baseline, at 3, 6, and 12 months follow-up.
Data will be compared between intervention and control group. Two (longitudinal) analyses
will be performed: 1) a crude analysis of outcome measures, and 2) an analysis of
outcome measures adjusted for prognostic dissimilarities, confounding, and effect
modification. Multilevel analyses will be performed to take into account repeated
measurements in participants as well as the effects of clustering in this study. Clusters
will be taken into account in this analysis as fixed effects. All analyses will be
performed according to the intention-to-treat principle. Multilevel analyses will
be performed using MlWin 2.0. Linear and logistic regression analyses will be performed
using SPSS.

Economic evaluation

An economic evaluation will be performed to measure and analyse the cost differences
between the current implementation strategy and passive dissemination of the guideline
in relation to the effect differences between the two groups. Direct costs related
to health care utility as well as indirect costs related to productivity losses due
to absenteeism and presenteeism will be analysed. Costs of the implementation strategy
will be measured, valued and analysed using a bottom-up approach. Health care consumption
and productivity losses will be valued using Dutch guideline prices.

Cost-effectiveness and cost-utility analyses will be performed from a societal perspective.
Functional status and quality of life of the patients will be taken as outcome measures
for the economic evaluation. To calculate incremental cost-effectiveness ratios (ICER)
and cost-utility ratios, the differences in mean costs between the two groups will
be divided by their differences in outcome measures. Confidence intervals (95 %) will
be estimated using bootstrapping methods with a minimum of 5.000 replications 23]. Uncertainty of the ICERs will be graphically presented in a cost-effectiveness plane
(CE plane), as well as cost-effectiveness acceptability curves (CEAC) 24], 25].

At last, a budget impact analysis will extrapolate the results of the economic evaluation
to a period of 5 years in order to estimate the financial costs of applying this implementation
strategy on a larger scale. This analysis will be performed from several perspectives
(i.e., societal, government, and insurance) and several scenarios for application
of this implementation strategy.