Int. J. Environ. Res. Public Health 2020, 17, 3942; doi:10.3390/ijerph17113942 www.mdpi.com/journal/ijerph
Patterning in Patient Referral to and Uptake of a
National Exercise Referral Scheme (NERS) in Wales
From 2008 to 2017: A Data Linkage Study
*, Muhammad Rahman
and Graham Moore
Centre for Development, Evaluation, Complexity and Implementation in Public Health Improvement
(DECIPHer), School of Social Sciences, Cardiff University, 1-3 Museum Place, Cardiff CF10 3BD, UK;
Swansea University Medical School, Swansea University, Swansea SA2 8PP, UK;
* Correspondence: Morgank22@cardiff.ac.uk; Tel.: +44-(0)-2920-870296
Received: 15 May 2020; Accepted: 30 May 2020; Published: 2 June 2020
Abstract: Exercise referral schemes have shown small but positive impacts in randomized
controlled trials (RCTs). Less is known about the long-term reach of scaled up schemes following a
RCT. A RCT of the National Exercise Referral Scheme (NERS) in Wales was completed in 2010, and
the scheme scaled up across Wales. In this study, using a retrospective data linkage design,
anonymized NERS data were linked to routine health records for referrals between 2008 and 2017.
Rates of referral and uptake were modelled across years and a multilevel logistic regression model
examined predictors of uptake. In total, 83,598 patients have been referred to the scheme and 67.31%
of eligible patients took up NERS. Older adults and referrals for a musculoskeletal or level four
condition were more likely to take up NERS. Males, mental health referrals, non-GP referrals and
those in the most deprived groupings were less likely to take up NERS. Trends revealed an overall
decrease over time in referrals and uptake rates among the most deprived grouping relative to those
in the least deprived group. Findings indicate a widening of inequality in referral and uptake
following positive RCT findings, both in terms of patient socioeconomic status and referrals for
Keywords: physical activity; public health; intervention; exercise referral; primary care
Globally it is estimated that 1.4 billion adults are insufficiently active, with one in four exercising
for less than 30 min five times a week . Such levels of inactivity date back to the 1990s  and today
women and adults from high income countries are shown to be among the least active . Inactivity
has been shown to increase the risk of developing diabetes, certain cancer types and heart disease,
and it is currently the fourth leading risk factor for mortality . As such, increasing physical activity
at the population level, and amongst at-risk populations, is a global public health priority .
A wide range of universal approaches have been employed to increase physical activity levels,
both at the individual and population level. One such approach originating in the early 1990s and
continuing to increase in popularity is the use of an Exercise Referral Scheme (ERS) [5,6]. Essentially,
an ERS involves the referral of individuals deemed to be insufficiently inactive from a health
practitioner to a structured exercise programme which is led by a third party.
The number of existing ERS is thought to be in excess of 600 schemes in the UK. However,
evidence concerning their effectiveness remains mixed . A recent review  including six
Randomised Controlled Trials (RCTs) noted positive impacts among patients referred for
cardiovascular or mental health reasons, yet a lack of evidence for musculoskeletal referrals. Weak
Int. J. Environ. Res. Public Health 2020, 17, 3942 2 of 16
evidence was found for ERS less than 12-weeks in duration and scheme uptake and adherence were
highlighted as inherent determinants of ERS success. Despite the basis of sound theory and well-
designed studies, evidence-based interventions are not being adopted widely or implemented with
sufficient quality and fidelity [9,10]. As such, it is recognised that RCTs are only the beginning of a
process which may lead to better public health outcomes ; ultimately an intervention, in this case
an ERS, needs to be adopted, implemented and sustained in order to have the greatest impact.
The number of individuals who are referred to any one scheme remains unknown, with studies
typically reporting on those patients who take up a scheme only. Understanding the volume of
patients referred to an ERS and, more so, the characteristics of those referred is essential in order to
begin to interpret the generalisability of findings to the wider population. These data would also
provide an insight into the accessibility or ERS, ascertaining whether schemes are being delivered
equitably as stipulated in the National Quality Assurance Framework for Exercise Referral . Some
have highlighted the potential for ERS to have an inequitable impact [13,14], with resources being
allocated and misspent on individuals who are ’worried well’, that is individuals who would have
taken up exercise anyway. Furthermore, concerns surrounding the ability of leisure sectors to attract
and retain individuals who are most likely to benefit from an active lifestyle have also been raised
After a patient is referred to an ERS, the patient may or may not then take up the scheme. Scheme
uptake has commonly been defined as a patient attending an initial consultation; however, this
definition is variable and at times lacking throughout the existing literature. Historically, patient
uptake rates have been reported at around 65% [15–17], with individual study reports ranging
between 35%  to 85%  and higher rates apparent among RCTs, suggesting participants in RCTs
may exhibit higher scheme engagement than the population as a whole where an intervention is
delivered outside of a RCT context. In cases where participants in RCTs are not reflective of the
intended wider population, for example, with higher motivation levels or younger ages, the external
validity of a RCT to the wider population is threatened  and as such the effectiveness of an
intervention within the real-setting remains unknown. A review of reviews  recently noted that
no progress has been made in increasing the number of participants starting ERS over the years since
the emergence of ERS. As Glasgow et al. argue , for an intervention to achieve population health
benefits it needs to be both efficacious for those who participate and to achieve wide reach through
its target population, as well as being widely adopted, implemented and maintained by the systems
in which it is intended to be delivered. Moving forward, there is a need for thoughtful approaches to
clinical evidence generation, supplementing RCT evidence with data generated from observational
studies for a greater understanding of what happens beyond a trial or study setting. Understanding
factors associated with patient uptake can help inform decision making processes of scheme referrers
and deliverers to help optimise uptake rates across patient sub-groups. Review findings suggest that
women [17,23] and older aged adults  are more likely to uptake ERS yet some individual studies
have found no patterning by patient gender [24,25] or age [26,27]. Evidence surrounding uptake rates
and patient socioeconomic status is currently deemed insufficient to determine any conclusion .
One review suggests reason for referral as a predictor of scheme uptake , with those referred for
a specific medical condition more likely to uptake, while other reviews report less clear findings .
A recent qualitative study  offers some insight into factors which influence uptake decisions,
highlighting factors across the intrapersonal (past physical activity experience, motivation,
competing demands), interpersonal (support and scheme understanding) and organisational
(promotion, communication and cost) levels of the socioecological model. As portrayed above, the
current evidence base appears somewhat conflicting, which is a likely reflection of the adoption of
different uptake definitions and, even more so, the notable divergence in ERSs both within and across
countries . Furthermore, to our knowledge, no study to date has reported detailed characteristics
of participants who are referred but later fail to take up an ERS.
Ascertaining ways in which scheme uptake can be maximised has been identified as a key
challenge for future research . To address this challenge, there is a need to improve our
understanding of the characteristics of referred patients in addition to the factors which predict
Int. J. Environ. Res. Public Health 2020, 17, 3942 3 of 16
patient uptake. Such information is vital in order to inform the current use of ERSs across the world
and to help address the levels of physical inactivity. One avenue is to utilise existing data
infrastructures and capitalise on patient data which is routinely collected by an ERS, a
recommendation within the 2014 National Institute for Health and Care Excellence (NICE) guidelines
. To date, only one database has been cited in the UK which contains longitudinal ERS data on
approximately 24,000 individuals across 19 schemes . Given the fragmented ERS system across
the UK, however, this database reflects the sporadic nature of local schemes and as such the
heterogeneity in scheme design skews early findings.
In Wales, UK, the National Exercise Referral Scheme (NERS) offers a unique opportunity to
harness readily available patient-level data to facilitate evaluative methods and inform evidence-
based policymaking. NERS is a standardised, evidence-based ERS which continues to be delivered
across all regions of Wales following positive findings from a RCT conducted from 2007–2010 .
Since its inception in 2007, the scheme has continued to collate data at the local level, gathering data
on each patient referred to the scheme. Specifically, data are collated from the point of scheme referral
through to a 12-month post-scheme follow-up. Given the growing international infrastructure
dedicated to facilitating data linkage  and the high level of granularity in data collected by NERS
over the years, this dataset has the potential to considerably influence policy and enhance our
understanding of patient journeys across an ERS, and to understand what happens with regards to
scheme participation once an intervention moves beyond the confines of a RCT and into every day
practice. Importantly, in the trial of NERS, improved outcomes were observed only among patients
who engaged fully with the intervention, with uptake and completion of the scheme offering useful
though imprecise proxies for effectiveness. If scheme engagement has increased since the trial, this
perhaps provides some indirect evidence that its population level impacts are likely greater than
suggested by the RCT. Conversely, if engagement has declined in routine practice, effectiveness is
also under threat.
The present study will utilise the National Exercise Referral scheme (NERS) as a case study for
addressing the above needs. Specifically, for all patients referred to NERS via a generic pathway, this
study aims to perform record linkage between NERS data and routine health records, model rates of
patient referral between 2008 and 2017 and understand factors associated with scheme uptake.
2. Materials and Methods
2.1. Setting and Participants
The NERS intervention is delivered by all of the 22 local authorities in Wales, in a variety of
settings including council owned leisure centers and private gyms. For referral, patients must be aged
16 years or above, be sedentary (not moderately active for 3 times per week/deconditioned through
age or inactivity), and have at least one of the following: raised blood pressure over 140/90, Body
Mass Index (BMI) over 28, cholesterol over 5.0, controlled diabetes or impaired glucose intolerance,
family history of heart disease or diabetes, at risk of osteoporosis or musculoskeletal pain, mild
arthritis or poor mobility, mild-moderate COPD (Chronic obstructive pulmonary disease), mild
anxiety or depression, multiple sclerosis. All patients referred via the generic pathway to the NERS
intervention between 13 February 2008 to 31 December 2017 were eligible to be included in this study.
2.2. NERS Intervention
The NERS intervention has been described previously in detail . Briefly, the scheme involves
the referral of a patient to an exercise professional and a subsequent 16-week programme of
subsidized, supervised group-based activity. Patients are predominantly referred via a General
Practitioner (GP) or physiotherapist. During the programme there is an initial consultation and a
follow-up contact at 4-weeks, 16-weeks and at 12-months, with all data recorded on a national
database by each local area. Following completion of the programme, patients are signposted to
alternative physical opportunities (i.e. ‘exit routes’), which typically remain within the leisure setting.
Int. J. Environ. Res. Public Health 2020, 17, 3942 4 of 16
2.3. Study Data
For each patient referred to NERS, data are routinely collected and recorded within the NERS
national database, provided by scheme referrers and scheme deliverers. Data held within the NERS
National database have been anonymised and deposited into the Secure Anonymous Information
Linkage (SAIL) databank , a well-established health-informatics which houses a wide array of
health and non-health datasets. The anonymization process uses a split-file approach whereby the
demographic data and clinical data for each patient are separated by the data provider and sent to a
third party provider. The third party assigns a unique linkage field (ALF) to each patient which
allows the split files to be recombined while removing any identifiable data during the process .
For the purpose of the current study and wider research project, the NERS national database
was linked to the following datasets: Welsh Demographic Service (WDS), Primary Care Dataset
(GPD), Patient Episode Database Wales (PEDW), Outpatient dataset (OPD), and the Annual District
Death Extract (ADDE). That is, for each patient within the NERS national database, scheme data were
linked to the clinical records within each of the aforementioned datasets. Before analyses, the linked
database was checked for consistency and errors. Data relating specifically to the current study are
outlined as follows.
Patient age (scheme referral date minus date of birth) and gender (Male/Female/Unknown) were
extracted from the GPD. For each patient, a deprivation score was assigned at the Lower Super
Output Area (LSOA) using the 2014 Welsh Index of Multiple Deprivation (WIMD) . WIMD 2014
comprises eight domains of social deprivation whereby each LSOA is ranked by an overall score and
categorised into quintiles. Quintile 1 represents the 20% most deprived LSOAs and quintile 5 the 20%
least deprived LSOAs. Further detail on the domains and score calculations is available elsewhere
Patient referral date, the occupation of the referring health professional and the reason for
patient referral were provided by the NERS national database. Referring health professionals (23
categories) were recoded as General Practitioner (GP; including practice nurse referrals),
physiotherapist or other. Reasons for referral (121 codes) were recoded as Coronary Heart Disease
(CHD) only, mental health only, CHD and mental health, musculoskeletal, and level four conditions.
Level four conditions signify patients deemed to be at ‘higher risk’ and who need to undertake
tailored exercise sessions as part of their rehabilitation following an NHS (National Health Service)
intervention or to manage a chronic condition and use physical activity as a means of secondary
Each local authority was categorised according to their involvement in the earlier RCT as trial
(1) or non-trial (0) area. An aggregated deprivation score was assigned to each local authority, based
on the percentage of LSOAs within each local authority which are ranked in the most deprived 50%
of LSOAs in Wales .
Within the present study, a referral indicated an action by a referring health professional while
uptake was defined as a patient attending a first consultation. Reasons for not taking up the scheme
were recorded within the NERS database (193 codes) which were recoded into 10 sub-categories:
intrinsic factors, scheme specifics, health-related, unknown, other commitments, exercising
elsewhere, work-related, transferred area/GP, bereavement, and accessibility.
2.4. Statistical Analysis
All quantitative statistical analyses were conducted in Stata v.16（StataCorp. 2019. College
Station, TX: StataCorp LLC). Descriptive analyses were performed to describe participant
characteristics (i.e. age, gender and socioeconomic status) across the whole sample and each uptake
group. Rates of referral and scheme uptake were modelled over time, with accompanying confidence
intervals to reveal any increase or decrease in rates across years. To assess whether patient
socioeconomic levels have changed over time, interaction effects were examined between the year of
referral and LSOA quintile.
As data were hierarchical in nature and individuals nested within local authorities, a multilevel
modelling approach was adopted. Specifically, a multi-level mixed-effects logistic regression model
Int. J. Environ. Res. Public Health 2020, 17, 3942 5 of 16
was used to examine predictors of scheme uptake, comparing characteristics of patients who took up
the scheme (uptake) to those who did not (non-uptake). Individual-level predictors included age,
gender, socioeconomic status, referral reason, referring professional and year of referral. Area-level
predictors included trial area (yes or no) and area deprivation. ‘Uptake’ was set as the reference
category and odds ratios (ORs) were reported alongside 95% confidence intervals (95%CI) and
accompanying p values. Differences were considered statistically significant if the p value was <0.05.
Ethical approval for this study was obtained from Cardiff University’s School of Social Sciences
Research Ethics committee (SREC/2163). This study was approved by the Health Information
Research Unit’s independent Information Governance Review Panel (project 0670).
3.1. Data Linkage
The provision of identifiable data enabled 93.58% (n = 168,977) of patient records to be
successfully anonymised. Of these, 78.61% (n = 141,955) were deterministically linked. For 8668
records (containing 2 duplicate records) within the NERS database, data linkage was not possible.
After the removal of duplicate patient linkages, 143,170 patients remained. Next, duplicate records
within the NERS national database (N = 19) and patients who had been referred on more than one
occasion (N = 9849) were discounted, creating a final dataset of 83,598 patients. Figure 1 provides an
overview of the process in creating a linked dataset, whereby the final dataset does not contain any
Figure 1. Overview of data linkage process.
Int. J. Environ. Res. Public Health 2020, 17, 3942 6 of 16
3.2. Scheme Referral
Since 2007, a total of 83,598 patients with linked data have been referred to NERS via the generic
pathway on one occasion. Patients were aged between 9 and 99 years (mean 53, SD 16.6) and
predominantly women (61.6%). The age profile of women (52.33 years, SD 16.53, range 9–99 years)
and men (55.54 years, SD 16.65, range 12–97 years) was comparable. Table 1 shows the proportion of
referrals across deprivation quintiles and that the most common reason for referral was
musculoskeletal related (50.4%) and most common referrer a GP (57.8%). Similar proportions were
noted among NERS referrals where data linkage was not possible (67.24% GP referrals and 43.23%
musculoskeletal related referrals; Supplementary Table S1).
Table 1. Characteristics of all National Exercise Referral Scheme (NERS) referrals between 2008 and
5 (Least deprived)
1 (Most deprived)
Reason for referral
Mental health only
CHD and mental health
* does not include patients with repeat referrals.
Figure 2 displays overall referral rates across each year and presents patients referred as a
percentage of the overall inactive population in Wales (based on waves of the National Survey for
Wales ). The graph displays referrals between 2011 and 2017 only as preceding data did not
provide a true picture of the total referrals. In total 3.3% of the population have been referred to NERS
since 2011. As shown, the percentage of referred patients gradually increased with a peak during
2013 before decreasing to a steady plateau between 2014 to 2017. Figure 3 provides a breakdown of
referrals over the years according to socioeconomic grouping. Trends reveal an overall decrease in
referrals of patients from the most deprived grouping alongside an increase in referrals of those in
the least deprived group. Following the divergence in referral rates between these two groups over
time, comparable rates are shown in 2017.
Int. J. Environ. Res. Public Health 2020, 17, 3942 7 of 16
Figure 2. Generic pathway referrals shown as a percentage of the total inactive population in Wales
(N = 83,598 patients).
Figure 3. Generic pathway referrals according to socioeconomic groups between 2011 and 2017 (5 =
lowest deprivation, 1 = highest deprivation) (N = 76,628 patients).
3.3. Scheme Uptake
Patients on a waiting list (7487), under age 16 years (n = 48) or classified as an inappropriate
referral (n = 593) were removed from analyses. The sample comprised a total of 75,470 patients, of
which 61.65% were women and 60.44% were aged 59 years or younger (range 16–99 years). As shown
in Table 2, patient deprivation scores indicate that 40.70% of the population were in the top two
deprived quintiles. Over half of patients (57.96%) had been referred by a GP, with musculoskeletal
referrals recorded as the most common referral reason (50.53%). Mental health related referrals
accounted for approximately one fifth of all patients (20.72%).
2011 2012 2013 2014 2015 2016 2017
% of population referred
% of referrals % of inactive (33%)
2011 2012 2013 2014 2015 2016 2017
% of patients
5 4 3 2 1
Int. J. Environ. Res. Public Health 2020, 17, 3942 8 of 16
Table 2. Characteristics of patients according to scheme uptake (N = 75,470).
Did Not Take Up
28,873 38.26 19,229 (66.60) 9644 (33.40)
5 (Least deprived)
13,721 19.79 9594 (69.92) 4127 (30.08)
1 (Most deprived)
Reason for referral
Mental health only
CHD and mental health
Year of referral ^
^ Years 2008–2010 statistical disclosure with numbers reflecting magnitude.
In total, 50,800 (67.31%) patients took up the scheme with rates across each year displayed in
Figure 4A. As shown, rates of uptake were highest in 2010 (above 80%) and declining thereafter to
2013 (63.58%). A slight increase in rates is shown between 2013 and 2017 (69.51%). Figure 4B shows
uptakes rates according to a patient’s deprivation quintile. For patients within the least deprived
group (quintile 5), uptake rates have steadily increased since 2010. For those among the most
deprived groups (quintiles 1 and 2), a continual decrease in uptake rates was found since 2009, and
in 2017 the lowest uptake was among those in the most deprived quintile.
Int. J. Environ. Res. Public Health 2020, 17, 3942 9 of 16
Figure 4. (A) Percentage of generic pathway patients taking up NERS across 2009 to 2017 (95%CI) (N
= 75,470 patients). (B) Percentage of generic pathway patients up taking NERS across 2009 to 2017
according to deprivation quintile (5 = least deprivation, 1 = most deprived) (N = 69,316 patients).
Reasons for not taking up the scheme were unknown for the majority of patients (83.51%) while
health-related (3.84%), other commitments (3.32%) and exercising elsewhere (3.36%) were the next
commonly recorded reasons.
3.4. Predictors of Scheme Uptake
Results from multilevel logistic regression analyses (Table 3) revealed that patients taking up
the scheme were more likely to be older (aged 45–59: OR 1.50: 95%CI:1.44–1.56; aged 60 plus: OR 2.37:
95%CI:2.27–2.47) or referred for a musculoskeletal (OR 1.15: 95%CI:1.09–1.20) or level four condition
(OR 1.90: 95%CI:1.76–2.05) in comparison to patients not taking up the scheme. Contrastingly,
2009 2010 2011 2012 2013 2014 2015 2016 2017
% of patients
Upper CI % Lower CI
2009 2010 2011 2012 2013 2014 2015 2016 2017
5 4 3 2 1
Int. J. Environ. Res. Public Health 2020, 17, 3942 10 of 16
patients who were male (OR 0.90: 95%CI:0.87–0.94), referred for mental health only (OR 0.79:
95%CI:0.74–0.84) or referred by a physiotherapist (OR 0.88: 95%CI:0.85–0.92) or a professional outside
of a GP surgery (OR 0.84: 95%CI:0.80–0.89) were found to be less likely to take up the scheme. No
area level variables were found to be associated with patient scheme uptake. As shown, a downward
linear trend was observed between deprivation quintile and uptake, with patients less likely to
uptake the scheme as the deprivation level increased.
Table 3. Multilevel logistic regression outputs of correlates to scheme uptake (N = 69,291).
Characteristic OR p
95% Confidence Interval
Age (16–44 years)
Deprivation quintile (5-least deprived)
1 (most deprived)
Reason for referral (CHD only)
Mental health only
CHD and mental health
Referral year (2017)
2009 3.15 0.006 1.39 7.14
Trial area (Yes)
ICC—level 1 variables
ICC—level 1 & 2 variables
Bold values signify significant findings p < 0.05.
This study has performed the most comprehensive assessment yet of referrals to and uptake of
an ERS. In doing so, this is the first study to create an electronic cohort of ERS patients, whereby
routinely collected ERS data have been linked to routine healthcare records. As such, data are now
available for wider research in an anonymized format and patients referred to NERS can be both
retrospectively and prospectively followed up within the SAIL databank.
Between 2008 and 2017, over 83,500 patients have been referred to NERS via the generic referral
pathway and since linked to routine data. Not taking into account those with existing comorbidities,
this suggests that approximately 3.3% of the ‘at risk’ inactive population  have been referred over
the 10-year period. Over this time, we observed a peak in referrals during 2013 which would reflect
Int. J. Environ. Res. Public Health 2020, 17, 3942 11 of 16
a time point at which NERS fully operated in all 22 local authorities in Wales. Thereafter, we found a
decline in referrals followed by a plateau in numbers in recent years. One possible explanation for
this observation relates to the expansion of NERS since the earlier RCT. Given the evidence-base for
generic referrals  and increasing demand from wider healthcare professionals, trial areas began
to increase the number of referral pathways available, with up to 13 currently running in some areas.
As such, the expansion of alternate referral pathways, alongside no increase in funding or capacity
for delivery, is a likely explanation for a drop in the number of generic referrals.
Of those referred to NERS, over two thirds of patients have taken up the scheme, which is
somewhat significantly lower than the earlier RCT  (at 85%) yet it is reflective of other ERS uptake
rates [15–17]. This finding enhances our understanding of what happens beyond a trial, denoting ERS
uptake rates in the real-world setting. This perhaps indicates that no matter how pragmatic, RCTs
represent a particular context in which to implement an intervention, which may not be wholly
reflective of real world practice, and hence there is a need for careful monitoring in real world practice
beyond a RCT . Effects of NERS for those patients who took it up might have been similar to those
observed in the trial but declines in uptake perhaps indicate that average effects per referred patient
are likely to have slightly reduced.
In contrast to our observed pattern of referral rates, uptakes rate showed a decline from 2010 to
2013, but a plateauing or marginal steady increase since. The initial spike of uptake rates in 2011 could
reflect uptake among a group of highly motivated individuals following the immediate period after
the RCT. Earlier interviews indicated that some GPs were hesitant to refer patients until
randomisation was removed , while the scheme was likely eagerly awaited in areas joining after
the trial, leading to high initial demand. While reviewing reviews, Shore et al.,  recently noted
that no progress has been made in increasing the number of participants starting a scheme over the
years. As ours is the first study to model national trends within the same ERS, the stable rate of uptake
over the past five years is somewhat encouraging, portraying the continued demand ten years on
from scheme inception. Similar to other studies, we noted that older adults  and women [17,23]
were more likely to take up the scheme, while patients referred via a physiotherapist or alternate
health professional were found to be less likely to take up the scheme compared to GP referrals. Given
the important role that health professionals play in facilitating change in physical activity , the
latter finding provides insight into the varying roles and influences referrers may have. Previous
studies have shown significant individual inter-clinician variability on how physical activity is
prescribed among patient groups , while Yarborough and colleagues noted key clinician traits for
helping patients with mental illnesses initiate and maintain lifestyle changes . Findings may also
reflect patient characteristics, with those from highly-educated or highly-motivated groups seeking
a referral via a GP . Wider literature has indicated that the type of referring professional may be
important for patient adherence to ERS .
Several studies [42–44], including the earlier RCT , have shown the positive impacts of ERS
for individuals suffering from anxiety and depression, with improved mental health outcomes post-
scheme completion. Moreover, above and beyond the ERS literature, exercise is recommended as an
evidence-based treatment for depression, the first and only mental health disorder whereby exercise
is recommended in clinical guidelines . In the present study, however, we found that patients
referred for a mental health reason were substantially less likely to take up the scheme compared to
those with a CHD referral. This finding reflects earlier reports from two other UK-based studies
[40,46]. With one in six individuals suffering from a mental health disorder  and approximately
two-thirds of individuals with a known mental illness not seeking help from a health professional
, major depressive disorder is predicted to become the leading cause of disability in the world by
2030 . Of all referrals to NERS, approximately 11% were dedicated to mental health and of these,
just over half of patients took up the scheme, representing a large decline in uptake among this group
compared to levels observed in the trial period. It is therefore important for future work to
understand the reasons underpinning the observed findings and, more so, vital to identify potential
facilitators to improve referral rates and uptake among this sub-group.
Int. J. Environ. Res. Public Health 2020, 17, 3942 12 of 16
Historically, an objective of local ERSs has been to improve health inequalities . Findings of
the present study showed that patients are being referred across the full spectrum of socioeconomic
circumstances, a finding which is consistent with the earlier RCT. As most chronic conditions and
physical activity levels are patterned by socioeconomic status , the need for ERS will be greater
among those in deprived groups compared to more affluent peers. That said, we noted a clear
downward trend for the referral of patients in the most deprived groupings over time, coupled with
a marked pattern of disparity in uptake rates, with patients in the most deprived groups least likely
to take up the scheme. Of note, the rapid decline in uptake rates among the most deprived groups
coincides with a change to the scheme pricing. In 2011, following the RCT, the scheme saw an increase
in cost from £1.00 to £1.50 per session. Thereafter, between 2013 and 2014 local authorities were
informed that scheme prices could be determined locally before once again becoming standardized
across all areas at £2 per session in 2014. Trends observed over the years are indicative of a growing
inequality—as uptake rates of the most affluent groups are continuing to increase, uptake rates of the
most deprived groups are continuing to decrease. Our observations are supportive of the inverse care
law, whereby patients considered most likely in need of health care are the least likely to receive the
care . While the trial reported no difference in effects by socioeconomic status, these data suggest
that in the period following the trial, it is likely that the intervention has become increasingly
inequality generating. Implicating factors could involve the length of waiting lists in areas and also
factors at the referral level. Drawing comparisons between health services within deprived and
affluent communities, studies have noted shorter GP appointment times  and fewer GP staff in
deprived areas . Only one other study has reported on scheme uptake according to patient
deprivation status. Harrison et al.,  found that patients who were more deprived and suffered
from a respiratory diagnosis were more likely to take up ERS compared to more affluent patients
with the same diagnoses. Further research is needed to understand the reasons underpinning the
observed patterning of referrals and uptake so that NERS can ensure that future services are
continued to be equitably available and accessible.
The totality of evidence presented in the current study can be used as a guide for both scheme
referrers and deliverers in identifying the best referral approaches across patient sub-groups. It could
be argued that different approaches need to be adopted according to patient characteristics and
current referral methods and follow-on processes need to be reviewed. A key strength of this study
is the population approach, involving over 85,000 patients and data spanning a ten-year period. This
is also the first study to examine an ERS following a RCT and national scale-up, therefore findings
are reflective of an established scheme. Our measures of socioeconomic status incorporated both
individual- and area-level data, reducing the likelihood of encountering the problem of ecological
fallacy . That said, our study lacks data on patient ethnicity, levels of physical activity and cases
where patients have requested a referral (i.e. ‘self-referral’). Such information would have provided
a greater insight into sub-groups and our observed trends. While we noted that 3.3% of patients
reported alternate exercise as a reason for not taking up the scheme, poor data capture prevents us
from analyzing this data further. In addition, the coding criteria used by health professionals to define
a referral health condition limits the ability to study differing degrees of health conditions (for
example, whether mild anxiety and/or depression).
As Wales is faced with the largest aging population of all UK nations  and levels of inactivity
persist, research into NERS continues to be as important as ever. The projected rise in the number of
over 70s in the coming years  is likely to place considerable pressure on the scheme and wider
public services. Future research will continue to follow each patient along the NERS journey, to
examine rates of scheme completion and patient drop-out, along with associated characteristics.
Analyses are currently underway to uncover any longer-term impacts of the scheme on patient health
outcomes and qualitative work is set to provide invaluable insight into the current findings.
The creation of an e-cohort allows us to maximize opportunities to conduct longitudinal research
and uncover any longer-term effects of participation in NERS. That said, our findings raise questions
Int. J. Environ. Res. Public Health 2020, 17, 3942 13 of 16
about the equity of NERS with diverging trends among referrals and uptake rates of socioeconomic
groups. Having uncovered different patterns of uptake according to patient sub-groups, findings can
be used by scheme referrers and deliverers to inform future processes.
Supplementary Materials: The following are available online at www.mdpi.com/1660-4601/17/11/3942/s1, Table
S1: Records of NERS referrals with no data linkage (N = 8666).
Author Contributions: Conceptualization, K.M.; methodology, K.M and G.M.; software, M.R.; formal analysis,
K.M.; data curation, K.M and M.R.; writing—original draft preparation, K.M.; writing—review and editing, K.M
and G.M.; funding acquisition, K.M. All authors have read and agreed to the published version of the
Funding: This work was funded by a Health and Care Research Wales Health Fellowship Award [grant number
HF-16-1164 to K.M.]. The Centre for Development, Evaluation, Complexity and Implementation in Public Health
Improvement (DECIPHer) is funded by Welsh Government through Health and Care Research Wales. This work
was supported also by the Public Health Division, Welsh Government with the support of The Centre for the
Development and Evaluation of Complex Interventions for Public Health Improvement (DECIPHer), a UKCRC
Public Health Research Centre of Excellence. Joint funding [grant number MR/KO232331/1] from the British
Heart Foundation, Cancer Research UK, Economic and Social Research Council, Medical Research Council, the
Welsh Government, and the Wellcome Trust, under the auspices of the UK Clinical Research Collaboration, is
Acknowledgments: The authors would like to thank the National Exercise Referral Scheme (NERS) delivery
team for their support during this study and Welsh Local Government Association for their assistance with data
curation. We would also like to thank Dan Thayer and Daniel Mallory for their assistance with acquiring WIMD
data. We are thankful for the invaluable input and guidance from our Public Involvement Community, Karen
Shepherd and Alan Meudell.
Conflicts of Interest: G.M. was part of the team who undertook the earlier RCT of the National Exercise Referral
Scheme in Wales. Other authors declare no conflict of interest. The funders had no role in the design of the study;
in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish
1. Guthold, R.; Stevens, G.A.; Riley, L.M.; Bull, F.C. Worldwide trends in insufficient physical activity from
2001 to 2016: A pooled analysis of 358 population-based surveys with 1.9 million participants. Lancet Glob.
Health 2018, 6, e1077–e86.
2. Haskell, W.; Lee, I.-M.; Pate, R.; Powell, K.; Blair, S.; Franklin, B.; Macera, C.; Heath, G.; Thompson, P.;
Bauman, A. Physical activity and public health: Updated recommendation for adults from the American
College of Sports Medicine and the American Heart Association. J. Circ. 2007, 16, 1081–93.
3. World Health Organisation. Global Recommendations on Physical Activity for Health; World Health
Organisation: Geneva, Switzerland, 2010.
4. Global action plan on physical activity 2018–2030: More active people for a healthier world. World Health
Organisation: Geneva, Switzerland, 2018.
5. National Institute for Health and Clinical Excellence. A Rapid Review of the Effectiveness of Exercise
Referral Schemes to Promote Physical Activity in Adults. National Institute for Health and Clinical
Excellence: London, UK, 2006.
6. Pavey, T.G.; Taylor, A.H.; Fox, K.R.; Hillsdon, M.; Anokye, N.; Campbell, J.L.; Foster, C.; Green, C.;
Moxham, T.; Mutrie, N. et al. Effect of exercise referral schemes in primary care on physical activity and
improving health outcomes: Systematic review and meta-analysis. BMJ 2011, 343, d6462.
7. Wadem, M.; Mann, S.; Copeland, R.J.; Steele, J. Effect of exercise referral schemes upon health and well-
being: Initial observational insights using individual patient data meta-analysis from the National Referral
Database. J. Epidemiol. Commun. Health 2020, 74, 32–41.
8. Rowley, N.; Mann, S.; Steele, J.; Horton, E.; Jimenez, A. The effects of exercise referral schemes in the United
Kingdom in those with cardiovascular, mental health, and musculoskeletal disorders: A preliminary
systematic review. BM Public Health 2018, 18, 949.
Int. J. Environ. Res. Public Health 2020, 17, 3942 14 of 16
9. Glasgow, R.E.; Lichtenstein, E.; Marcus, A.C. Why don't we see more translation of health promotion
research to practice? Rethinking the efficacy-to-effectiveness transition. Am. J. Public Health 2003, 93, 1261–
10. Chaudoir, S.R.; Dugan, A.G.; Barr, C.H. Measuring factors affecting implementation of health innovations:
A systematic review of structural, organizational, provider, patient, and innovation level measures.
Implement Sci. 2013, 8, 22.
11. Bumbarger, B.; Perkins, D. After randomised trials: Issues related to dissemination of evidence-based
interventions. J. Child. Serv. 2008, 3, 55–64.
12. Department of Health. Exercise Referral Systems: A National Quality Assurance Framework. Department
of Health: London, UK, 2001.
13. Iliffe, S.T.S.; Gould, M.; Thorogood, M.; Hillsdon, M. Prescribing exercise in general practice. BMJ 1994,
14. Hillsdon, M.T.M. A systematic review of physical activity promotion strategies. Br. J. Sports Med. 1996, 30,
15. Pavey, T.G.; Anokye, N.; Taylor, A.H.; Trueman, P.; Moxham, T.; Fox, K.R.; Hillsdon, M.; Green, C.;
Campbell, J.L.; Foster, C.; et al. The clinical effectiveness and cost-effectiveness of exercise referral schemes:
A systematic review and economic evaluation. Health Technol Assess. 2011, 15,1–254.
16. Williams, N.H.; Hendry, M.; France, B.; Lewis, R.; Wilkinson, C. Effectiveness of exercise-referral schemes
to promote physical activity in adults: Systematic review. Br. J. Gener. Pract.: J. R. Coll. Gener. Practiti. 2007,
17. Gidlow, C.; Johnston, L.H.; Crone, D.; James, D. Attendance of exercise referral schemes in the UK: A
systematic review. Health Educ. J. 2005, 64, 168–86.
18. Stevens, W.; Hillsdon, M.; Thorogood, M.; McArdle, D. Cost-effectiveness of a primary care based physical
activity intervention in 45-74 year old men and women: A randomised controlled trial. Br. J. Sports Med.
1998, 32, 236–41.
19. Murphy, S.M.; Edwards, R.T.; Williams, N.; Raisanen, L.; Moore, G.; Linck, P.; Hounsome N.; Din, N.;
Moore, L. An evaluation of the effectiveness and cost effectiveness of the National Exercise Referral Scheme
in Wales, UK: A randomised controlled trial of a public health policy initiative. J. Epidemiol. Commun. Health
2012, 66, 745–53.
20. Kennedy-Martin, T.; Curtis, S.; Faries, D.; Robinson, S.; Johnston, J. A literature review on the
representativeness of randomized controlled trial samples and implications for the external validity of trial
results. Trials 2015, 16, 495.
21. Shore, C.B.; Hubbard, G.; Gorely, T.; Polson, R.; Hunter, A.; Galloway, S.D. Insufficient reporting of factors
associated with exercise referral scheme uptake, attendance, and adherence: A systematic review of
reviews. J. Phys. Act. Health 2019, 16, 667.
22. Glasgow, R.E.; Vogt, T.M.; Boles, S.M. Evaluating the public health impact of health promotion
interventions: The RE-AIM framework. Am. J. Public Health 1999, 89, 1322–7.
23. Pavey, T.; Taylor, A.; Hillsdon, M.; Fox, K.; Campbell, J.; Foster, C.; Moxham, T.; Mutrie, N.; Searle, J.;
Taylor, R. Levels and predictors of exercise referral scheme uptake and adherence: A systematic review. J.
Epidemiol. Commun. Health 2012, 66, 737–44.
24. Harrison, R.A.; McNair, F.; Dugdill, L. Access to exercise referral schemes—A population based analysis.
J. Public Health 2005, 27, 326–30.
25. Gidlow, C.; Johnston, L.H.; Crone, D.; Morris, C.; Smith, A.; Foster, C.; James, D. Socio-demographic
patterning of referral, uptake and attendance in Physical Activity Referral Schemes. J. Public Health 2007,
26. Isaacs, A.J.; Critchley, J.A.; Tai, S.S, Buckingham, K.; Westley, D.; Harridge, S.D.; Smith, C.; Gottlieb, J.
Exercise Evaluation Randomised Trial (EXERT): A randomised trial comparing GP referral for leisure
centre-based exercise, community-based walking and advice only. Health Technol Assess. 2007, 11, 1–165,
27. Damush, T.M.; Stump, T.E.; Saporito, A.; Clark, D.O. Predictors of older primary care patients' participation
in a submaximal exercise test and a supervised, low-impact exercise class. Prevent. Med. 2001, 33, 485–94.
28. Birtwistle, S.B.; Ashcroft, G.; Murphy, R.; Gee, I.; Poole, H.; Watson, P.M. Factors influencing patient uptake
of an exercise referral scheme: A qualitative study. Health Educ. Res. 2018, 34, 113–27.
Int. J. Environ. Res. Public Health 2020, 17, 3942 15 of 16
29. National Institute for Health and Care Excellence. Physical activity: Exercise referral schemes. Public health
guideline [PH54]. 2014. Available online: https://www.nice.org.uk/guidance/ph54 (accessed on 24
30. Steele, J.; Wade, M.; Polley, M.; Copeland, R.J.; Stokes, S.; Mann, S. The National Referral Database: An
initial overview. Available online: https://doi.org/10.31236/osf.io/rgywq. 2019 (accessed on 21 March 2020).
31. Jones, K.; Ford, D.V.; Thompson, S.; Lyons, R.A. A profile of the SAIL databank on the UK secure research
platform. Int. J. Popu. Data Sci. 2019, 4, 03.
32. Jones, K.H.; Ford, D.V.; Jones, C.; Dsilva, R.; Thompson, S.; Brooks, C.J.; et al. A case study of the Secure
Anonymous Information Linkage (SAIL) Gateway: A privacy-protecting remote access system for health-
related research and evaluation. J. Biomed. Inform. 2014, 50, 196–204.
33. Welsh Government. Welsh Index of Multiple Deprivation 2014. Executive summary. Available online:
wimd-2014-executive-summary.pdf. 2014 (accessed on 9 August 2018).
34. Statistics for Wales. National Survey for Wales 2018-19: Adult lifestyle. 2019. Available online: (accessed on
16 July 2019).
35. Lehmann, B.A.; Lindert, L.; Ohlmeier, S.; Schlomann, L.; Pfaff, H.; Choi, K.-E. “And then he got into the
wrong group”: A qualitative study exploring the effects of randomization in recruitment to a randomized
controlled trial. Int. J. Environ. Res. Public Health 2020, 17, 1886.
36. Din, N.U.; Moore, G.F.; Murphy, S.; Wilkinson, C.; Williams, N.H. Health professionals’ perspectives on
exercise referral and physical activity promotion in primary care: Findings from a process evaluation of the
National Exercise Referral Scheme in Wales. Health Educ. J. 2015, 74, 743–57.
37. NICE. Physical Activity: Brief Advice for Adults in Primary Care. NICE: London, UK, 2013.
38. Hansen, D.; Rovelo Ruiz, G.; Doherty, P.; Iliou, M.C.; Vromen, T.; Hinton, S.; Frederix, I.; Wilhelm, M.;
Schmid, J.; Abreu, A.; et al. Do clinicians prescribe exercise similarly in patients with different
cardiovascular diseases? Findings from the EAPC EXPERT working group survey. European J. Prevent.
Cardiol. 2018, 25, 682–91.
39. Yarborough, B.J.H.; Stumbo, S.P.; Cavese, J.A.; Yarborough, M.T.; Green, C.A. Patient perspectives on how
living with a mental illness affects making and maintaining healthy lifestyle changes. Patient Educ. Counsel.
2019, 102, 346–51.
40. Sowden, S.L.; Breeze, E.; Barber, J.; Raine, R. Do general practices provide equitable access to physical
activity interventions? Br. J. Gener. Pract. 2008, 58, e1–e8.
41. Dugdill, L.; Graham, R.C.; McNair, F. Exercise referral: The public health panacea for physical activity
promotion? A critical perspective of exercise referral schemes; their development and evaluation.
Ergonomics 2005, 48, 1390–410.
42. Duda, J.L.; Williams, G.C.; Ntoumanis, N.; Daley, A.; Eves, F.F.; Mutrie, N.; Rouse, P.; Lodhia, R.; Blamey,R.;
Jolly, K. Effects of a standard provision versus an autonomy supportive exercise referral programme on
physical activity, quality of life and well-being indicators: A cluster randomised controlled trial. Int. J.
Behav. Nutr. Phys. Acti. 2014, 11, 10.
43. Rouse, P.C.; Ntoumanis, N.; Duda, J.L.; Jolly, K.; Williams, G.C. In the beginning: Role of autonomy support
on the motivation, mental health and intentions of participants entering an exercise referral scheme.
Psychol. Health 2011, 26, 729-49.
44. Chalder, M.; Wiles, N.J.; Campbell, J.; Hollinghurst, S.P.; Haase, A.M.; Taylor, A.H.; Fox, K.; Costelloe, C.;
Searle, A.; Baxter, H.; Winder, R.; et al. Facilitated physical activity as a treatment for depressed adults:
Randomised controlled trial. Bmj: Br. Med. J. 2012, 344, e2758.
45. Glowacki, K.; Arbour-Nicitopoulos, K.; Burrows, M.; Chesick, L.; Heinemann, L.; Irving, S.; Lam, R.;
Macridis, S.; Michalak, E.; Scott, A.; et al. It's more than just a referral: Development of an evidence-
informed exercise and depression toolkit. Mental Health Phys. Acti. 2019, 17, 100297.
46. Crone, D.; Johnston, L.H.; Gidlow, C.; Henley, C.; James, D.V. Uptake and participation in physical activity
referral schemes in the UK: An investigation of patients referred with mental health problems. Issues Ment
Health Nurs. 2008, 29, 1088-97.
47. National Institute for Health and Care Excellence. NICEimpact mental health. NICE: London, UK, 2019.
48. Department of Health. Annual Report of the Chief Medical Officer 2013, Public Mental Health Priorities:
Investing in the Evidence. Available online: Gov.uk/government/publications/chief-medical-officer-
cmoannual-report-public-mental-health. 2014 (accessed on 19 February 2016).
Int. J. Environ. Res. Public Health 2020, 17, 3942 16 of 16
49. Yang, L.; Zhao, Y.; Wang, Y.; Liu, L.; Zhang, X.; Li, B.; Cui, R. The effects of psychological stress on
depression. Curr. Neuropharmacol. 2015, 13, 494-504.
50. Marmot, M.W.R. Social Determinants of Health. 2nd ed.; Oxford University Press: Oxford, UK, 2006.
51. Tudor Hart, J. The inverse care law. Lancet 1971, 297, 405-12.
52. Mercer, S.W.; Watt, G.C. The inverse care law: Clinical primary care encounters in deprived and affluent
areas of Scotland. Anna. Fam. Med. 2007, 5, 503-10.
53. Dayan, M. Are parts of England 'left behind' by the NHS? Nuffield Trust comment. Available online:
https://www.nuffieldtrust.org.uk/news-item/are-parts-of-england-left-behind-by-the-nhs 2018 (accessed
on 11 May 2019).
54. MacRae, K. Commentary: Socioeconomic deprivation and health and the ecological fallacy. BMJ 1994, 309,
55. Age Discrimination Info. Age discrimination statistics. Available online:
http://www.agediscrimination.info/statistics 2020 (accessed on 22 Janurary 2020).
56. National Assembly for Wales Finance Committee. The cost of caring for an ageing population. Available
online: https://www.assembly.wales/laid%20documents/cr-ld11773/cr-ld11773-e.pdf. National Assembly
for Wales Commission; 2018 (accessed on 22 January 2020).
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).