Preventive Medicine 84 (2016) 48–56

Contents lists available at ScienceDirect

Preventive Medicine

journal homepage: www.elsevier.com/locate/ypmed

A community based primary prevention programme for type 2 diabetes

integrating identification and lifestyle intervention for prevention: the
Let's Prevent Diabetes cluster randomised controlled trial☆
Melanie J. Davies a,b, Laura J. Gray c,⁎, Jacqui Troughton d, Alastair Gray e, Jaakko Tuomilehto f,g,h, Azhar Farooqi a,
Kamlesh Khunti a, Thomas Yates a,b, on behalf of the Let's Prevent Diabetes Team:
a
Diabetes Research Centre, University of Leicester, Leicester, UK
b
Leicester–Loughborough Diet, Lifestyle and Physical Activity Biomedical Research Unit, Leicester, UK
c
Department of Health Sciences, University of Leicester, Leicester, UK
d
Leicester Diabetes Centre, University Hospitals of Leicester, Leicester, UK
e
Health Economics Research Centre, Nuffield Department of Population Health, University of Oxford, Oxford, UK
f
Centre for Vascular Prevention, Danube University Krems, 3500 Krems, Austria
g
Department of Chronic Disease Prevention, National Institute for Health and Welfare, 00271 Helsinki, Finland
h
Diabetes Research Group, King Abdulaziz University, 21589 Jeddah, Saudi Arabia

a r t i c l e i n f o a b s t r a c t

Available online 29 December 2015 Objectives. Prevention of type 2 diabetes (T2DM) is a priority in healthcare, but there is a lack of evidence
investigating how to effectively translate prevention research into a UK primary care setting. We assessed
Keywords: whether a structured education programme targeting lifestyle and behaviour change was effective at preventing
Type 2 diabetes progression to T2DM in people with pre-diabetes.
Pre-diabetes Materials and methods. Forty-four general practices were randomised to receive either standard care or a
Prevention
6 hour group structured education programme with an annual refresher course, and regular phone contact.
Lifestyle
Clinical trial
Participants were followed up for 3 years. The primary outcome was progression to T2DM.
Results. Eight hundred and eighty participants were included (36% female, mean age 64 years, 16% ethnic
minority group); 131 participants developed T2DM. There was a non-significant 26% reduced risk of developing
T2DM in the intervention arm compared to standard care (HR 0.74, 95% CI 0.48, 1.14, p = 0.18). The reduction in
T2DM risk when excluding those who did not attend the initial education session was also non-significant (HR
0.65, 0.41, 1.03, p = 0.07). There were statistically significant improvements in HbA1c (− 0.06, − 0.11,
− 0.01), LDL cholesterol (− 0.08, − 0.15, − 0.01), sedentary time (− 26.29, − 45.26, − 7.32) and step count
(498.15, 162.10, 834.20) when data were analysed across all time points.
Conclusions. This study suggests that a relatively low resource, pragmatic diabetes prevention programme
resulted in modest benefits to biomedical, lifestyle and psychosocial outcomes, however the reduction to the
risk of T2DM did not reach significance. The findings have important implications for future research and primary
care.
© 2016 Elsevier Inc. All rights reserved.

Introduction T2DM may be shortened by as much as 10 years, with most dying of car-
diovascular diseases (CVD) (Roper et al., 2001). The management of
Type 2 diabetes mellitus (T2DM) is associated with reduced quality T2DM consumes around 10% of health care expenditure (Hex et al.,
of life and serious complications. The life expectancy of individuals with 2012). Consequently, the prevention of T2DM is a priority and has
been highlighted by the NHS, UK, as one of four priority areas (NHS,
2014).
☆ Trial Registration: ISRCTN80605705. Pre-diabetes (PDM) is a high-risk state where glucose levels are ele-
⁎ Corresponding author at: Leicester Diabetes Centre (Bloom), Leicester General vated but do not reach the threshold for diagnosis of T2DM. Trials have
Hospital, Gwendolen Road, Leicester LE5 4PW, UK. unequivocally demonstrated that lifestyle interventions, which pro-
E-mail addresses: melanie.davies@uhl-tr.nhs.uk (M.J. Davies), lg48@le.ac.uk
(L.J. Gray), Jacqui.Troughton@uhl-tr.nhs.uk (J. Troughton), alastair.gray@dph.ox.ac.uk
mote moderate to vigorous-intensity physical activity, a healthy diet
(A. Gray), jaakko.tuomilehto@thl.fi (J. Tuomilehto), azhar_farooqi@hotmail.com and weight regulation, reduce the risk of progressing to T2DM by
(A. Farooqi), kk22@le.ac.uk (K. Khunti), ty20@le.ac.uk (T. Yates). 30%–60% in those with PDM (Gillies et al., 2007). For example, the

http://dx.doi.org/10.1016/j.ypmed.2015.12.012
0091-7435/© 2016 Elsevier Inc. All rights reserved.
 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56 49

Finnish Diabetes Prevention Study (DPS) found that the risk of T2DM Interventions
was reduced by 58% in those referred to an intensive lifestyle interven-
tion compared to usual care over a three-year period (Tuomilehto et al., All participants received an information booklet which included informa-
2001). Consistent findings have been reported from the USA Diabetes tion on risk factors for T2DM, and how dietary and lifestyle changes and in-
creased physical activity can prevent progression to T2DM.
Prevention Program (DPP) (Knowler et al., 2002).
The participants in the intervention practices were invited to attend the
Despite the strong evidence for lifestyle interventions in the pre-
Let's Prevent programme (Gray et al., 2012c), which tailors the widely delivered
vention of T2DM, there has been a translational gap between trial ev- DESMOND structured-education programme into a prevention context (Davies
idence and implementation into routine care. This is predominantly et al., 2008; Gillett et al., 2010).
due to the resource-intensive nature of lifestyle interventions tested. Let's Prevent was delivered to groups of ten over 6 h, either over a full-
For example, in the first year of the DPP programme, participants re- day or two half-days, by two trained educators. The programme was
ceived 16 1 h one-to-one counselling sessions followed by an aver- underpinned by a theoretical basis with a philosophy centred on patient
age of eight additional contacts and two telephone consultations. empowerment. The aim was to increase knowledge and promote realistic
The participants were also offered supervised exercise classes perceptions of PDM, and to promote healthy behaviour, with the aims of re-
(Knowler et al., 2002). This intensity of care is incompatible with ducing body weight by 5%, limiting total and saturated fat intake to 30% and
10% of total energy intake respectively, increasing fibre intake and promot-
routine care pathways. Therefore, the emphasis needs to be shifted
ing physical activity. The physical activity section incorporated the success-
to examining the effectiveness of approaches designed for imple-
ful PREPARE structured-education programme (Yates et al., 2009), based on
mentation within routine primary care. As healthcare services have providing participants with a pedometer and enabling the formation of
differences in funding, organisation and infrastructure, programmes personalised step-per-day goals. The content and educational resources
cannot be assumed to be generalisable across contexts. To date there used within the programme were further tailored to the need of local
has been a dearth of evidence concerning T2DM prevention in the South Asian populations, including delivery through interpreters where re-
UK, with small-scale projects showing mixed results (Yates et al., quired. The educators were trained using an accredited pathway, and re-
2009; Dyson et al., 1997; Oldroyd et al., 2006; Bhopal et al., 2014). ceived ongoing support and quality development to ensure consistent
This study assesses whether the Let's Prevent T2DM programme is delivery.
effective at preventing progression to T2DM in people with PDM identi- The participants were invited to 3 h refresher sessions at 12 and 24 months
fied through a systematic screening pathway within primary care. Let's to reinforce key messages, review risk factors and update action plans. In addi-
tion, the participants received a 15-minute telephone call every 3 months from
Prevent is a pragmatic, relatively low resource, group-based structured-
healthcare professionals trained to offer ongoing support in behaviour change.
education programme targeting lifestyle behaviour change specifically
Those who did not attend the initial session were not invited to the refresher
designed for implementation within a community setting. sessions, but continued to be followed up.

Methods
Outcome measures

The study had two phases. The first was a screening phase which identified
All outcomes were measured at the participant level. The primary out-
people at risk of PDM/T2DM through the use of a screening tool that had been
come was progression to T2DM during 3 years. T2DM diagnosis was defined
developed and validated for use within primary care (Gray et al., 2012a; Gray
according to WHO 1999 criteria/guidelines (World Health Organization,
et al., 2012b). In the second phase, the participants who had been screened
1999), Participants without symptoms of diabetes in whom the initial
and found to have PDM progressed to the cluster RCT. The cluster RCT design
OGTT showed T2DM were recalled for a second test to confirm the diagnosis.
has been described in detail elsewhere (Gray et al., 2012c). The trial randomised
Participants found to have T2DM at baseline were excluded. Following the
practices to avoid the risk of contamination. Ethical approval was sought and
update of the WHO diagnostic criteria to include HbA1c (World Health
the study involved practice level and individual level informed consent. The re-
Organization, 2011) we obtained a protocol amendment in January 2013
cruitment took place between May 2009 and June 2011, with follow-up data
allowing HbA1c ≥ 6.5% to become part of the diagnostic criteria for T2DM
collected up to July 2014.
within this study. Therefore T2DM was diagnosed using OGTT prior to Janu-
ary 2013, and with either an OGTT or HbA1c post January 2013. The partic-
ipants and their GP were informed of the results. The diagnosis of T2DM
Practices and participants
within primary care was also captured by self-report followed by confirma-
tion through GP records. Participants diagnosed with T2DM after baseline
Practices in Leicestershire, UK, were recruited and randomised using a remained in the study to complete the questionnaires and other biomedical
computer-generated list 1:1 to either the standard-care or intervention arm data, but did not undertake further OGTTs.
by an independent researcher, using stratification by list size (b 6000, ≥6000), A full list of the secondary outcomes assessed at each time point is de-
and ethnicity (percentage South Asian b 21%, ≥21% — taken from ADDITION- scribed elsewhere (Gray et al., 2012c), these included: lipid levels, HbA1c,
Leicester study; Webb et al., 2010) with a block size of four. Practices and partic- medical and medication history, blood pressure, weight, waist and body
ipants were informed of their allocation in the result letters after the screening/ mass index (BMI). The participants also completed a questionnaire contain-
baseline measurements were complete. Eligible participants were identified ing a number of validated questionnaires which assessed total self-reported
from recruited practices via a two-stage screening process. The Leicester Diabe- physical activity, subsequently reported as metabolic equivalent minutes
tes Practice Risk Score was used in each practice to identify people at high-risk per week (METS.mins/week) (Craig et al., 2003), diet reported as a unit-
of PDM/T2DM for invitation to screening (Gray et al., 2012a). The top 10% of pa- less fibre, total fat and unsaturated fat score (Roe et al., 1994), illness beliefs
tients with the highest score fulfilling the inclusion criteria were invited. The in- (Broadbent et al., 2006), anxiety and depression (Zigmond and Snaith,
clusion criteria for screening were ages 40 to 75 if White European, or 25– 2006), quality of life (Sintonen and Pekurinen, 1993) and sleep pattern; re-
75 years if South Asian. Participants were excluded if they were unable to give source usage data and EQ-5D responses were also collected for the cost-
informed consent, pregnant or lactating, had established diabetes or a terminal effectiveness analysis (Gusi et al., 2010). The participants also wore a sealed
illness, or if they required an interpreter for a language other than one of the lo- pedometer (NL-800, New Lifestyles, Inc., Lees Summit, MO, USA) with a
cally used South Asian languages accommodated within this study. All those seven-day memory during waking hours to provide habitual ambulatory ac-
agreeing to take part received an oral glucose tolerance test (OGTT). Only partic- tivity (average daily step count derived by summing total accumulated
ipants who were identified as having PDM (IFG and/or IGT WHO 1999 criteria; steps and dividing by days worn). For the purposes of this study, at least
World Health Organization, 1999) during screening took part in the RCT. In one three valid days of data were required; a valid day constituted at least
small practice (list size = 1650) no participants were identified with PDM and 10 h of wear time (Tudor-Locke and Bassett, 2004).
this practice was excluded. Other secondary outcomes included change in cardiovascular risk as calcu-
The screening–visit data formed the baseline assessment for the RCT; the lated by the Framingham risk calculator, and presence of metabolic syndrome
participants were followed up at 6, 12, 24 and 36 months. as defined by NCEP ATP III.
50 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56

Sample size higher in the standard-care group compared to the intervention

group.
The sample size takes into account the clustering by GP practice. Assuming a
three-year cumulative conversion rate to T2DM of 35% in the standard-care
Development of T2DM
group (equivalent to 117 events per 1000 follow-up years), an intraclass corre-
lation of 0.05, and 17 participants per practice (equal clusters assumed), we
131 participants developed T2DM during a median follow-up of
needed 280 per group to detect a 40% risk reduction in the intervention
group. Allowing for a 25% drop-out rate, a total of 748 participants needed to 3 years; this equates to 60.32 events per 1000 person years (95% CI
be recruited. 50.82, 71.58). Lower rates were seen in the intervention group: 57.60
events per 1000 person years compared to 63.16 events per 1000 per-
Statistical analysis son years in the standard-care group (Figure A.1). The hazard ratio
(HR) showed a non-significant 26% reduced risk of developing T2DM
A statistical-analysis plan was agreed before data were available. Practice in the intervention arm compared to standard-care (p = 0.18). The ef-
and participant level characteristics were compared by group, using either fect was greater (35% reduction) in the per-protocol analysis albeit
means (SD) or medians (IQR) for continuous variables, and counts and percent- still non-significant (p = 0.07) (Table 2). The risk of developing T2DM
ages for nominal variables. Cluster randomisation gives balance with respect to in the intervention group compared to standard-care was similar across
cluster-level covariates but can lead to imbalance in participant level covariates;
all sub-groups of PDM. The ICC for the development of T2DM was 0.02
therefore differences were assessed using t-tests and chi-squared tests.
(95% CI 0, 0.05).
Progression to T2DM was analysed on an intention-to-treat basis (ITT). The
event rate per 1000 person years was calculated by group. The participants were
censored at the date of their last clinical appointment or at diagnosis of T2DM. Secondary outcomes
Cox proportional hazard models with group as a covariate were fitted; practices
were assumed to have the same frailty. Hazard ratios (HR) and 95% confidence In both groups improvements were seen for many secondary out-
intervals were presented. Subgroup analyses were performed by PDM status comes, see Table 3 for a summary and A.3–A.6 for the full results. A sta-
(IGT, IFG, IGT and IFG, HbA1c 6.0%–6.4%). tistically significant reduction of 0.06% in HbA1c was seen in the
For all secondary outcomes, participants who developed T2DM during the intervention group compared to the standard-care group when
study had their last value from before their diagnosis carried forward for the re- analysing the mean across all time points. Significant reductions in
mainder of the study. This method was used in a previous similar study (Yates
LDL-cholesterol were seen at 12 months and overall. For all other out-
et al., 2009). All secondary outcomes were analysed using generalised estimat-
comes, apart from systolic blood pressure, a greater improvement was
ing equation models with an exchangeable correlation structure, which adjust-
seen in the intervention group than the standard-care group at
ed for clustering. For binary outcomes we used a logit link with a binomial
distribution for the outcome, and for continuous outcomes we used an identity 36 months, but none of these reached statistical significance. No differ-
link with a normal distribution. The analysis was repeated at each time point. ences were seen for CVD/CHD risk outcomes or the presence of meta-
The missing outcomes were not replaced and we derived an average of contin- bolic syndrome.
uous outcomes over time. This procedure measures the cumulative effect of the Improvements during the study were seen in illness perceptions,
intervention and has the maximum number of participants. quality of life and anxiety in the intervention group compared to the
Subsidiary analyses were carried out for the main secondary outcomes at 12 standard-care group. Fat intake was not significantly reduced in the in-
and 36 months. The analysis was repeated: (i) excluding those from the inter- tervention group compared to the standard-care group, but statistically
vention group who did not attend the initial Let's Prevent education session significant increases in unsaturated fat were reported. Although no dif-
(per-protocol); and (ii) imputing any missing values for the secondary out-
ferences between the intervention and standard-care group were seen
comes using multiple imputation (ITT). The imputation was carried out using
in the self-reported levels of activity, a significant reduction in sitting
the command MI in Stata. MI replaces missing values with multiple sets of sim-
ulated values to complete the data, performs standard analysis on each com-
time of 30 min per day or more was seen in the intervention group at
pleted dataset, and adjusts the obtained parameter estimates for missing-data 12 and 24 months and overall, compared to the standard-care group.
uncertainty using Rubin's rules to combine estimates.(Rubin, 2004) Adjust- There was also an increase in objectively measured average daily step-
ments were not made for multiple testing. All results from the planned analyses count in the intervention group of 450–500 steps per day at all time
have been reported and p values were interpreted taking into account the over- points, with a significant effect seen at 12, 36 months and overall.
all pattern of the results. Statistical significance was set at 5%. All analyses were
conducted using Stata version 13. Subsidiary analyses

Results In the per-protocol analysis, significant reductions at 36 months

were seen for fasting (− 0.12 mmol/l), 2-hour (− 0.35 mmol/l) and
Overall, 43 practices were included; ranging in size from 1650 to HbA1c (− 0.11%) in the intervention compared to standard-care
24,000. The median number of participants with PDM recruited per (Table 4). The increase in step count in the intervention group com-
practice was 23 in the standard-care arm and 17 in the intervention pared to standard-care was increased in the per-protocol analysis. For
arm; the number recruited per practice ranged from 2 to 49. In total the intention-to-treat analysis, the interpretation of the results for the
880 participants were recruited (433 standard care, 447 intervention, secondary outcomes assessed did not change.
Fig. 1). At 36 months 76% of the intervention group attended compared
to 79% in the standard-care arm (p = 0.43). Compared to those who Discussion
attended at 36 months, non-attendees were more likely to be smokers,
and from more socially-deprived locations (Table A.1). Of those partici- To our knowledge, this is the first study investigating the effective-
pants from practices randomised to the intervention, 101 (22.6%) did ness of a T2DM prevention programme within primary care in the UK.
not attend the initial education session and were excluded in per- We have shown that a pragmatic, low-resource, three-year T2DM
protocol analyses. Compared to those who attended, non-attendees prevention programme, based on a 6 h , group-based, structured-
were younger, more likely to be male, from more socially deprived loca- education session followed by two annual group-based sessions and
tions, more likely to be smokers, and less physically active (see nine telephone contacts, can lead to improvements in markers of meta-
Table A.2). bolic health, psychosocial well-being, and health behaviour. The prima-
At baseline higher levels of deprivation and smokers were seen in ry outcome of the study was reduction in progression to T2DM;
the intervention group (Table 1). Weight (p = 0.0002), BMI (p = although non-significant, a modest 25% reduction in progression was
0.003) and waist circumference (p = 0.0001) were significantly observed in those receiving the education intervention, which increased
 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56 51

Fig. 1. Flow of practices and participants through the trial.

to 35% when excluding those who did not attend the first education not statistically significant, the effect size was similar to the Indian Dia-
session. betes Prevention programme, which reported a 28.9% reduction in the
Our study has several strengths and limitations. The strengths in- risk of T2DM following a lifestyle intervention (Ramachandran et al.,
clude using a rigorous design to evaluate the effectiveness of the pro- 2006). Limitations inherent in cluster randomised studies were also ob-
gramme specially developed for delivery within a multiethnic primary served here, particularly achieving a balance in participant characteris-
care setting. The primary limitation was that this study was underpow- tics across groups: important differences at baseline were observed,
ered due to the discrepancy between predicted and observed incidence with the intervention group having higher levels of social deprivation
rates of T2DM. The observed incidence rate of 63.16 events per 1000 and smoking rates, but with lower levels of BMI and waist circumfer-
person years in the standard-care arm was substantially lower than an- ence. These differences could have acted to confound the result. Finally,
ticipated, and consistent with those observed in non-intervention set- the generalisability of the findings should be assessed cautiously. Of
tings (Morris et al., 2013). This acted to substantially reduce the study those invited for screening, 19% attended (Gray et al., 2012b). Although
power, resulting in wider confidence intervals and a greater likelihood this uptake rate is consistent with other studies in similar populations
of a type 2 error. In addition, the variation in cluster size was greater (Aujla et al., 2010; Webb et al., 2011) and reflects the difficulty of
than planned (the number of participants recruited per practice ranged recruiting a multiethnic urban population into prevention studies, it
from 2 to 49), further diluting power. However, the estimates of the ICC may limit the generalisability of the findings. Higher rates of uptake
(0.05) and the inflation for dropout (25%) used were adequate (0.02 would be expected in a non-research setting: for example, the NHS
and 24% respectively). Although the reduction in the risk of T2DM was Health Checks programme has 40% uptake (Dalton et al., 2011). Due
52 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56

Table 1 a priori number secondary outcomes (Gray et al., 2012c), and this re-
Baseline characteristics. flects the nature of the intervention, which targets a number of aspects
Data given as mean (SD) unless otherwise stated.
of health and well-being. We have not adjusted for multiple testing,
Standard care Intervention which may have increased the type 1 error rate (Freemantle, 2001).
Individual level The results seen here reflect those seen in other similar trials
(Ramachandran et al., 2006; Lindström et al., 2003), and have been
Number of participants 433 447
interpreted in terms of clinical as well as statistical importance.
Age 63.9 (7.9) 63.9 (7.6)
Our study extends evidence for efficacy of lifestyle interventions in
Male n (%) 278 (64.2) 282 (63.1)
White European, n (%) 363 (84.3) 377 (84.5)
the prevention of T2DM into in a primary care setting. There is now ev-
Deprivation, median (IQR) 10.1 (6.3, 18.1) 13.4 (8.4, 24.4)⁎ idence internationally that T2DM prevention programmes can be tai-
Current smoker, n (%) 22 (5.1) 38 (8.5)⁎ lored for, and translated into, primary health care and community
Prescribed statins, n (%) 171 (43.3) 184 (44.2) settings, with modest short-term effects on markers of health status,
Prescribed antihypertensives, n (%) 270 (62.4) 275 (61.5)
such as body weight (Dunkley et al., 2014). However, longer-term stud-
History CVD, n (%) 78 (18.0) 75 (16.8)
HbA1c (%) 6.1 (0.4) 6.1 (0.4) ies designed to quantify effectiveness on reducing progression to T2DM
HbA1c (mmol/mol) 42.8 (4.6) 43.2 (4.7) are lacking. This has resulted in a lack of evidence-based solutions that
Fasting glucose 5.6 (0.7) 5.7 (0.7) might enable primary care organisations to conform to NICE guidance
2-hour glucose 8.8 (1.6) 8.9 (1.7)
for the prevention of T2DM (National Institute of Health and Clinical
Total cholesterol (mmol/l) 5.1 (1.1) 5.0 (1.0)
HDL cholesterol (mmol/l) 1.4 (0.5) 1.4 (0.5)
Excellence, 2012). By utilising structured education, Let's Prevent was
LDL cholesterol (mmol/l) 3.0 (0.9) 3.0 (0.9) purposefully designed to harness existing infrastructure within routine
Triglycerides (mmol/l) 1.7 (1.0) 1.7 (0.9) primary care. Structured education has been recommended in the man-
Systolic blood pressure (mmHg) 147.7 (17.7) 147.9 (20.7) agement of T2DM by NICE since 2003 (National Institute for Health and
Diastolic blood pressure (mmHg) 86.2 (10.6) 86.6 (11.0)
Clinical Excellence, 2003). DESMOND is one of the most prominent na-
Heart rate (bmp) 69.1 (12.1) 68.3 (13.1)
Weight (kg) 94.4 (18.9) 89.9 (16.6)⁎ tionally available T2DM structured-education programme, and the only
BMI (kg/m2) 33.1 (5.8) 32.0 (5.2)⁎ UK programme tested within a multi-centred RCT to quantify effective-
Waist circumference (cm) 111.3 (13.2) 108.0 (12.4)⁎ ness and cost-effectiveness (Davies et al., 2008; Gillett et al., 2015). Here
Average steps per day 6308.12 (3094.44) 6137.97 (2791.02) we show that this approach can be adapted to the prevention of T2DM
IFG only, n (%) 51 (11.8) 57 (12.8)
IGT only, n (%) 308 (71.1) 301 (67.3)
within a diverse multiethnic PDM population whilst using less than 25%
IFG and IGT, n (%) 74 (17.1) 89 (19.9) of the contact time seen in other efficacy trials. Future research is need-
HbA1c 6.0%–6.4% (%) 216 (50.1) 205 (46.7) ed to investigate how the approach used in Let's Prevent can be tailored
CHD 10 year risk⁎⁎⁎ 14.5 (8.6) 14.5 (8.5) to individual preferences concerning the frequency and format of con-
CVD 10 year risk⁎⁎⁎ 20.2 (11.8) 20.3 (11.9)
tact. In particular, utilising web-based platforms is likely to receive a
Cluster level
Number of practices 20⁎⁎ 23 growing focus in the future.
Median participants per practice (IQR) 23 (8, 34) 17 (7, 30) A separate paper assesses the cost effectiveness of the intervention.
Range participants per practice 2–41 3–49 In brief, this showed that the education programme was associated
Median practice size (IQR) 6932 (4008, 10,069) 5429 (3356, 8780) with higher costs (£168) and higher quality of life (0.046 QALYs) com-
High South Asian population, n (%) 2 (10.0) 4 (17.4)
pared to the standard care group over 3 years. Therefore, the Let's Pre-
HDL: High-density lipoprotein; LDL: Low-density lipoprotein; IGT: Impaired glucose vent programme is likely to be cost effective at a willingness to pay a
tolerance; IFG: Impaired fasting glucose; CHD: Coronary heart disease; CVD:
threshold of £20,000 per QALY gained (Leal et al., 2015).
Cardiovascular disease; IQR: Inter-quartile range.
⁎ Groups differ significantly (p b 0.05).
⁎⁎ One practice randomised to standard care had no eligible participants. Conclusion
⁎⁎⁎ Only calculated for those of WE or SA ethnicity between the ages of 35 and 75.
We have shown that a relatively low-resource, pragmatic T2DM pre-
to ethical constraints, no data were extracted from primary care regard- vention programme can lead to modest improvements to biomedical,
ing those invited for screening. Therefore we cannot compare the char- lifestyle and psychosocial outcomes without significantly reducing the
acteristics of the invited cohort to those who attended to establish if risk of T2DM. The findings have important implications for future re-
there was any potential for bias. A study of similar design conducted search and primary care.
in the same area screening for T2DM found that the 22% who attended
the screening were older and more likely to be female compared to non- Funding and ethics
attenders (Webb et al., 2011). Only 77% of the intervention group
attended the initial education, ad-hoc analyses suggest that the inter- This research was funded by the National Institute for Health Re-
vention effect can be increased as attendance increases. Future studies search (NIHR) under its Programme Grants for Applied Research
should focus on strategies to increase uptake to screening and atten- scheme (RP-PG-0606-1272). This report/article presents independent
dance/compliance with the programme. research commissioned by the National Institute for Health Research
We observed several improvements in secondary outcomes. Impor- (NIHR) under its Programme Grants for Applied Research scheme
tantly, there was evidence that the programme reduced threatening (RP-PG-0606-1272). The views expressed in this publication are those
perceptions of PDM, anxiety and improved overall quality of life. This of the authors and not necessarily those of the NHS, the NIHR or the De-
is consistent with other structured-education programmes (Davies partment of Health. Ethical approval was granted for this project by the
et al., 2008). These benefits were mirrored by modest improvements NHS East Midlands Ethics Committee.
in health behaviour. For example, the intervention group reported
healthier dietary fat profile, sitting on average for 26 min less per day Contributor statements
and undertaking 498 more steps per day compared to standard-care.
This change equates to an increase of 35 min of purposeful walking MJD is the principal investigator for the Let's Prevent programme
per week (Tudor-Locke and Bassett, 2004). This is similar to that report- grant, initiated the project, commented on the drafts of the paper and
ed in the Early Activity in Diabetes (ACTID) diet and physical activity in- approved the final version. MJD is the guarantor for the paper and
tervention for those with newly diagnosed T2DM, which was also affirms that the manuscript is an honest, accurate, and transparent
conducted in primary care (Andrews et al., 2015). The protocol defined account of the study being reported, that no important aspects of the
 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56 53

Table 2
Development of T2DM overall and by PDM diagnosis.
HR (95% CI) takes into account clustering.

Standard care Intervention HR 95% CI p value

Intention to treata 0.74 0.48, 1.14 0.18

Events, n (%) 67 (15.5) 64 (14.3)
Rate per 1000 person years (95% CI) 63.16 (49.71, 80.24) 57.60 (45.09, 73.59)
Per protocol 0.65 0.41, 1.03 0.07
Events, n (%) 67 (15.5) 51 (14.7)
Rate per 1000 person years (95% CI) 63.16 (49.71, 80.24) 53.04 (40.31, 69.80)
PDM subgroup
IGT alone 34 (11.2) 32 (10.7) 0.79 0.45, 1.38 0.41
IFG alone 7 (14.3) 6 (10.5) 0.52 0.15, 1.83 0.31
IGT and IFG 26 (32.9) 26 (28.3) 0.51 0.22, 1.16 0.11
HbA1c 6.0–6.4 36 (15.2) 27 (11.9) 0.65 0.38, 1.12 0.12
a
This is the same as complete case as there are no missing data for the primary outcome.

study have been omitted, and that any discrepancies from the study as commented on the drafts of the paper and approved the final version.
planned (and, if relevant, registered) have been explained. LJG wrote AG had input into the design of the study, the collection of data, led
the statistical analysis plan, cleaned and analysed the data, and drafted the cost effectiveness analysis, commented on the drafts of the paper
and revised the paper. JT developed the Let's Prevent programme, and approved the final version. JTu had input into the design of the

Table 3
Secondary outcomes.
Coefficients show the mean effect of the intervention compared to standard care adjusted for baseline value and cluster (full table given in Tables A3–6).

6 months 12 months 24 months 36 months Overall

Fasting glucose NR 0.001 (−0.10, 0.10) −0.06 (−0.16, 0.04) −0.05 (−0.18, 0.07) 0.0004 (−0.10, 0.10)
2-hour glucose NR 0.08 (−0.23, 0.39) −0.07 (−0.37, 0.22) −0.14 (−0.46, 0.18) −0.03 (−0.28, 0.22)
HbA1c (%) −0.07 (−0.12, −0.04 (−0.10, 0.02) −0.10 (−0.20, −0.004)⁎ −0.07 (−0.18, 0.04) −0.06 (−0.11, −0.01)⁎
−0.01)⁎
Total cholesterol (mmol/l) −0.06 (−0.18, 0.05) −0.07 (−0.16, 0.02) −0.02 (−0.12, 0.08) −0.11 (−0.23, 0.02) −0.06 (−0.14, 0.01)
HDL cholesterol (mmol/l) 0.003 (−0.05, 0.06) −0.01 (−0.07, 0.05) 0.004 (−0.06, 0.07) −0.02 (−0.08, 0.05) 0.01 (−0.04, 0.05)
LDL cholesterol (mmol/l) −0.06 (−0.15, 0.04) −0.10 (−0.018, −0.02)⁎ −0.02 (−0.09, 0.05) −0.09 (−0.19, 0.01) −0.08 (−0.15, −0.01)⁎
Triglyceride (mmol/l) −0.01 (−0.16, 0.14) 0.05 (−0.05, 0.15) −0.05 (−0.15, 0.05) −0.06 (−0.17, 0.06) −0.001 (−0.08, 0.08)
Body weight (kg) −0.10 (−0.72, 0.51) −0.27 (−1.17, 0.63) −0.49 (−1.48, 0.50) −0.26 (−1.17, 0.65) −0.10 (−0.85, 0.66)
BMI (kg/m2) −0.03 (−0.24, 0.19) −0.11 (−0.42, 0.21) −0.14 (−0.50, 0.21) −0.05 (−0.38, 0.27) −0.02 (−0.28, 0.25)
Waist circumference (cm) −0.91 (−2.03, 0.20) −0.11 (−1.37, 1.15) −0.82 (−2.03, 0.40) −0.79 (−1.73, 0.14) −0.45 (−1.32, 0.42)
Systolic BP (mmHg) 1.17 (−1.45, 3.79) 1.22 (−0.85, 3.30) −1.26 (−3.79, 1.28) 0.55 (−2.09, 3.19) 0.81 (−0.97, 2.60)
Diastolic BP (mmHg) −0.22 (−1.90, 1.46) 0.80 (−0.66, 2.26) −0.37 (−1.92, 1.19) −0.49 (−2.15, 1.17) 0.24 (−0.82, 1.30)
Heart rate (bpm) −1.31 (−2.90, 0.28) −0.61 (−1.84, 0.61) −0.68 (−2.00, 0.65) −0.52 (−1.83, 0.78) −0.66 (−1.58, 0.27)
CHD 10-year risk −0.18 (−0.84, 0.48) −0.14 (−0.73, 0.45) −0.68 (−1.40, 0.04) −0.23 (−1.07, 0.60) −0.35 (−0.81, 0.11)
CVD 10-year risk 0.28 (−0.77, 1.32) 0.01 (−0.74, 0.76) −0.74 (−1.64, 0.15) 0.04 (−1.07, 1.14) −0.14 (−0.64, 0.35)
Metabolic syndromea NR 1.05 (0.78, 1.43) 0.81 (0.60, 1.09) 1.10 (0.83, 1.46) 1.10 (0.83, 1.46)
Illness perception score −1.46 (−3.13, 0.21) −2.06 (−4.03, −0.09)⁎ −2.47 (−4.16, −0.78)⁎⁎ −1.16 (−2.69, 0.37) −1.61 (−2.92, −0.30)⁎
(BIPQ)b
Quality of life score (15D)c 0.01 (−0.001, 0.01) 0.01 (−0.002, 0.02) 0.01 (−0.002, 0.02) 0.02 (0.01, 0.03)⁎⁎ 0.01 (0.001, 0.02)⁎
Anxiety score (HADS)d −0.21 (−0.57, 0.15) −0.40 (−0.77, −0.03)⁎ −0.09 (−0.40, 0.21) −0.11 (−0.44, 0.23) −0.28 (−0.54, −0.02)⁎
Depression score (HADS)d −0.08 (−0.42, 0.26) −0.34 (−0.81, 0.14) −0.09 (−0.45, 0.27) −0.05 (−0.44, 0.35) −0.21 (−0.57, 0.16)
Diet (DINE):
Fibre intake −1.69 (−4.68, 1.29) 0.97 (−1.27, 3.21) −1.64 (−4.68, 1.39) 1.53 (−0.94, 4.00) −1.01 (−3.11, 1.08)
Fat intake −1.41 (−4.60, 1.77) 0.45 (−2.62, 3.51) −0.55 (−4.04, 2.95) −3.60 (−7.52, 0.31) −0.72 (−2.92, 1.48)
Unsaturated fat intake 0.18 (−0.12, 0.48) 0.32 (0.05, 0.58)⁎ 0.50 (0.24, 0.76)⁎⁎⁎ 0.38 (0.12, 0.63)⁎⁎ 0.33 (0.15, 0.51)⁎⁎⁎
Subjective physical activity
(IPAQ):
352.71 (−570.24,
Total METS 1275.65) 447.31 (−220.84, 1115.46) 415.06 (−234.47, 1064.59) −19.82 (−568.05, 528.41) 428.37 (−175.19, 1031.93)
−27.26 (−63.34,
Sitting time (min) 8.83) −25.94 (−49.95, −1.92)⁎ −38.96 (−66.15, −11.78)⁎⁎ −20.15 (−43.91, 3.60) −26.29 (−45.26, −7.32)⁎⁎
Objective physical activity:
591.38 (63.61,
Average steps per day 1119.16)⁎ 551.76 (117.27, 986.25)⁎ 466.30 (−65.50, 998.10) 535.76 (12.71, 1058.81)⁎ 498.15 (162.10, 834.20)⁎⁎
Sleep:
Hours slept last night NR 0.04 (−0.15, 0.22) −0.05 (−0.18, 0.09) −0.10 (−0.26, 0.06) −0.05 (−0.18, 0.08)
Average hours asleep in 24 h NR 0.10 (−0.16, 0.35) −0.03 (−0.23, 0.17) 0.11 (−0.06, 0.27) 0.01 (−0.16, 0.18)

HDL: High-density lipoprotein; LDL: Low-density lipoprotein; BMI: Body mass index; IGT: Impaired glucose tolerance; IFG: Impaired fasting glucose; CHD: Coronary heart disease; CVD:
Cardiovascular disease; BIPQ: Brief Illness Perception Questionnaire; 15D: 15-dimensional; HADS: Hospital Anxiety and Depression Scale; DINE: Dietary Instrument for Nutrition Educa-
tion; IPAQ: International Physical Activity Questionnaire; METS: Metabolic equivalents.
a
Data shown OR (95% CI).
b
Higher score reflects a more threatening view of illness.
c
Higher score reflects better quality of life.
d
Higher score reflects worse levels of anxiety/depression.
⁎ p b 0.05.
⁎⁎ p b 0.01.
⁎⁎⁎ p b 0.001.
54 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56

Coefficients show intervention effect adjusted for baseline value and cluster. Complete case — data analysed according to randomised group; those with missing data excluded on a case-by-case basis. Per protocol — those randomised to the inter-
study, commented on the drafts of the paper and approved the final ver-

469.52 (29.47, 909.57)⁎

−0.48 (−1.34, 0.37)
−0.02 (−0.13, 0.08)

−0.55 (−1.50, 0.39)

−0.10 (−0.45, 0.25)
−0.07 (−0.17, 0.04)
sion. AF had input into the design of the study, commented on the drafts

1.10 (0.83, 1.46)

of the paper and approved the final version. KK and TY had input into
the design and running of the study, commented on the drafts of the

36 months
paper and approved the final version.
The Let's Prevent Diabetes team: Keith Abrams, University of Leices-
ter, Leicester. Dariush Ahrabian, University of Oxford. Sayjal Amin, Uni-
vention who did not attend the initial education are excluded. Intention to treat — data analysed according to randomised group, all randomised participants included. Missing data imputed using multiple imputations.

versity Hospitals of Leicester, Leicester. Mary Bancroft, Hockley Farm

Medical Practice. Janette Barnett, University Hospitals of Leicester,
Leicester. Hannah Berkeley, University Hospitals of Leicester, Leicester.

576.47 (110.37, 1042.56)⁎

Danielle Bodicoat, University of Leicester, Leicester. Michael Bonar, Uni-
−0.16 (−0.81, 0.49)
0.02 (−0.09, 0.13)

−0.12 (−0.96, 0.72)

0.10 (−0.22, 0.42)
−0.02 (−0.08, 0.04)
versity Hospitals of Leicester, Leicester. Louise Boyles, University Hospi-

1.05 (0.78, 1.43)

Intention to treat

tals of Leicester, Leicester. Paul Bray, University Hospitals of Leicester,

Leicester. Nichola Cairns, University Hospitals of Leicester, Leicester.
12 months

Sandra Campbell, University Hospitals of Leicester, Leicester. Marian

Carey, University Hospitals of Leicester, Leicester. Patrice Carter, Univer-
sity of Leicester, Leicester. Sudesna Chatterjee, University Hospitals of
Leicester, Leicester. Pauline Cowling, Sheffield Teaching Hospitals NHS
Foundation Trust, Sheffield. Carolyn Currie, University Hospitals of
Leicester, Leicester. Heather Daly, University Hospitals of Leicester,
634.27 (1419.41, 2665.56)⁎
−0.35 (−0.61, −0.09)⁎⁎

Leicester. Alison Dunkley, University of Leicester, Leicester. Sue Enright,

−0.12 (−0.23, −0.01)⁎

−0.11 (−0.21, −0.01)⁎

University Hospitals of Leicester, Leicester. Stephanie Goldby, University

−0.15 (−1.05, 0.75)
0.12 (−1.03, 1.28)
0.77 (0.56, 1.04)

Hospitals of Leicester, Leicester. Geri Gray, University Hospitals of

Leicester, Leicester. Colin Greaves, University of Exeter Medical School,
36 months

Exeter. Joe Henson, University Hospitals of Leicester, Leicester. Stephen

Hiles, University Hospitals of Leicester, Leicester. Sian Hill, University
Hospitals of Leicester, Leicester. Jayne Hill, University Hospitals of
Leicester, Leicester. Hannah Holdsworth, University Hospitals of Leices-
ter, Leicester. Rosie Horne, University of Leicester, Leicester. Zin Zin
Htike, University Hospitals of Leicester, Leicester. Shenaz Jamal, Univer-
777.48 (336.66, 1218.31)⁎⁎

sity Hospitals of Leicester, Leicester. Janet Jarvis, University Hospitals of

Leicester, Leicester. Carolyn Johnson, University Hospitals of Leicester,
−0.11 (−0.75, 0.53)
0.03 (−0.14, 0.08)

−0.02 (−0.77, 0.73)

0.03 (−0.30, 0.36)
−0.04 (−0.10, 0.02)

0.74 (0.52, 1.05)

Leicester. Janet Jones, University Hospitals of Leicester, Leicester.

Kenneth Jones, University Hospitals of Leicester, Leicester. Sabera
Per protocol

Khan, University Hospitals of Leicester, Leicester. Anita Khulpateea, Uni-

12 months

versity Hospitals of Leicester, Leicester. Jose Leal, University of Oxford.

Judith Leonard, University Hospitals of Leicester, Leicester. Hamidreza
Mani, University Hospitals of Leicester, Leicester. Lorraine Martin-
Stacey, University Hospitals of Leicester, Leicester. Val Morgan, Univer-
sity Hospitals of Leicester, Leicester. Frances Morris, University Hospitals
535.76 (12.71, 1058.81)⁎

of Leicester, Leicester. Samiul Mostafa, University Hospitals of Leicester,

−0.05 (−0.18, 0.07)

0.04 (−1.07, 1.14)

−0.14 (−0.46, 0.18)
−0.07 (−0.18, 0.04)
−0.23 (−1.07, 0.60)

Leicester. Alison Northern, University Hospitals of Leicester, Leicester.

1.10 (0.83, 1.46)

Kayleigh O'Brien, University Hospitals of Leicester, Leicester. Hersha

Patel, University Hospitals of Leicester, Leicester. Naina Patel, University
36 months

of Leicester, Leicester. Rachel Plummer, University Hospitals of Leices-

ter, Leicester. Sheila Porter, University Hospitals of Leicester, Leicester.
Mo Radia, University Hospitals of Leicester, Leicester. Kathryn Realf,
University Hospitals of Leicester, Leicester. Dean Richmond, University
Hospitals of Leicester, Leicester. Clare Russell, University of Leicester,
Leicester. Rebecca Saker, University Hospitals of Leicester, Leicester.
551.76 (117.27, 986.25)⁎

CHD: Coronary heart disease; CVD: Cardiovascular disease.

Jane Sennet, University Hospitals of Leicester, Leicester. David Sheppard,

0.001 (−0.10, 0.10)

0.01 (−0.74, 0.76)

0.08 (−0.23, 0.39)
−0.04 (−0.10, 0.02)
−0.14 (−0.73, 0.45)

Saffron Group Practice, Leicester. Rebecca Spong, University of Leicester,

1.05 (0.78, 1.43)

Leicester. Bernie Stribling, University Hospitals of Leicester, Leicester.

Complete case
Key secondary outcomes — subsidiary analysis.

Margaret Stone, University of Leicester, Leicester. Nick Taub, University

12 months

of Leicester, Leicester. David Webb, University of Leicester, Leicester.

Emma Wilmott, University Hospitals of Leicester, Leicester. Carolina
Wilson, University Hospitals of Leicester, Leicester. Panna Wilson, Uni-
versity Hospitals of Leicester, Leicester.
Participating Practices: Dr Y B Shah & Partners, Silverdale Medical
Average steps per day

Centre. Leicester Medical Group Aylestone Rd Medical Centre.

Metabolic syndrome

Leicester Medical Group, Walnut Street Medical Centre. Dr J M Fry

CHD 10-year risk
CVD 10-year risk
Fasting glucose
2-hour glucose

& Partner (Dr Clay), Rushey Mead Health Centre. Dr Patchett &
Partners, Groby Road Medical Centre. Leicester Medical Group,
HbA1c (%)

⁎ p b 0.05.
⁎⁎ p b 0.01.

Thurmaston Health centre. Dr Mojaria, Broadhurst Street Medical

Table 4

Practice. Dr Clay & Partners, The Banks Surgery. Dr Brunskill &

Partners, Pinfold Medical Practice. Dr Barlow & Partners, Quorn
 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56 55

Medical Centre. Dr Clay & Partners, Cottage Surgery. Dr Joshi (Dr References
Astles & Partners), Willowbrook Medical Centre. Dr G Singh Pasley Andrews RC, Cooper AR, Montgomery AA, Norcross AJ, Peters TJ, Sharp DJ, et al. Diet or
Road Health Centre. Dr Ryan & Partners, Woodbrook Medical Centre. diet plus physical activity versus usual care in patients with newly diagnosed type
Dr Lennox & Partner, St Matthews Medical Centre. Dr Prassad, 2 diabetes: the Early ACTID randomised controlled trial. Lancet.378(9786):129–139.
Aujla, N., Eborall, H.C., Stone, M., Taub, N., Davies, M.J., Khunti, K., 2010. Barriers to practice
Clarendon Park Rd Health Centre. Dr G C Ackerley & Partners (Dr and patient recruitment to primary care based diabetes screening studies. Diabet.
Bandrapalli), Beaumont Lodge previously known as Heatherbrook Med. 27 (Suppl. 1), 371.
(Astill Lodge) Surgery (also site at Baxters Close). Dr Trzcinski & Bhopal, R.S., Douglas, A., Wallia, S., et al., 2014. Effect of a lifestyle intervention on weight
change in South Asian individuals in the UK at high risk of type 2 diabetes: a family-
Partners, Markfield Medical Centre. Dr Lewis & Dr Patel, Whitwick
cluster randomised controlled trial. Lancet Diabetes Endocrinol. 2 (3), 218–227.
Health Centre. Dr Davenport & Partners, Newbold Verdon Medical Broadbent, E., Petrie, K.J., Main, J., Weinman, J., 2006. The brief illness perception
Practice. Dr Wilmott & Partners, Castle Mead Medical Centre. Dr questionnaire. J. Psychiatr. Res. 60, 631–637.
Azar Farooqi, East Leicester Medical Practice. Dr Wilson & Partners, Craig, C.L., Marshall, A.L., Sjöström, M., et al., 2003. International physical activity
questionnaire: 12-country reliability and validity. Med. Sci. Sports Exerc. 35,
The Old School Surgery. Dr Bennett & Partners, Market Harborough 1381–1395.
Medical Centre. Dr D A Nandha, Evington Medical Centre. Dr Maini Dalton, A.R.H., Bottle, A., Okoro, C., Majeed, A., Millett, C., 2011. Uptake of the NHS Health
& Dr Roshan, The Willows Medical Centre. Dr Davies & Partners, Checks programme in a deprived, culturally diverse setting: cross-sectional study.
J. Public Health (doi:10.1093).
Ashby Health Centre. Dr Palin (Dr Prideaux & partners), Bushloe Davies, M., Heller, S., Skinner, T., et al., 2008. Effectiveness of the diabetes education and
End Surgery. Dr Ghatora, Shepshed Health Centre. Dr Woods & Part- self management for ongoing and newly diagnosed (DESMOND) programme for peo-
ners, Hugglescote Surgery. Dr S Mansingh & Dr SK Dey, St Peters ple with newly diagnosed type 2 diabetes: cluster randomised controlled trial. Br.
Med. J. 336 (7642), 491–495.
Health Centre. The Practice Asquith, Asquith Surgery. Dr H Dunkley, A., Bodicoat, D.H., Greaves, C.J., et al., 2014. Diabetes prevention in the real
Mukadam, Fosse Medical Centre. The Practice Cross Street. Dr world: effectiveness of pragmatic lifestyle interventions for the prevention of type
Bhutani & Partners, Enderby Medical Centre. The Practice Rushey 2 diabetes and of the impact of adherence to guideline recommendations. A system-
atic review and meta-analysis. Diabetes Care 37, 922–933.
Mead. The Practice — Sayeed Medical Centre. Dr JC Reynolds & Part-
Dyson, P.A., Hammersley, M.S., Morris, R.J., Holman, R.R., Turner, R.C., 1997. The Fasting
ners (Dr Graham Johnson), The Wycliffe Medical Practice, Hyperglycaemia Study: II. Randomized controlled trial of reinforced healthy-living
Lutterworth Medical Centre. Dr Pathak/Dr Roshan, Hazelmere Medi- advice in subjects with increased but not diabetic fasting plasma glucose. Metabolism
46 (12 Suppl. 1), 50–55.
cal Centre. Dr B W Kinsella & Partners — Hockley Farm Heath & Social
Freemantle, N., 2001. Interpreting the results of secondary end points and subgroup anal-
Care Centre. Dr Shafi, Briton Street Surgery. Dr Shafi — Westcotes 1, yses in clinical trials: should we lock the crazy aunt in the attic? BMJ[Br. Med. J.] 322
Westcotes GP Surgery W1. Dr Shafi — Westcotes 2 Westcotes GP Sur- (7292), 989–991.
gery W2. Dr Panton & Partners Oakmeadow Surgery. Gillett, M., Dallosso, H.M., Dixon, S., et al., 2010. Delivering the diabetes education and self
management for ongoing and newly diagnosed (DESMOND) programme for people
with newly diagnosed type 2 diabetes: cost effectiveness analysis. BMJ 341, c4093.
Gillett M, Dallosso HM, Dixon S, Brennan A, Carey ME, Campbell MJ, et al. Delivering the
Conflicts of interests Diabetes Education and Self Management for Ongoing and Newly Diagnosed
(DESMOND) Programme for People with Newly Diagnosed Type 2 Diabetes: Cost Ef-
fectiveness Analysis 2010 2010-08-20 12:28:44.
Laura J Gray, Jacqui Troughton, Alastair Gray, Jaakko Tuomilehto and Gillies, C.L., Abrams, K.R., Lambert, P.C., et al., 2007. Pharmacological and lifestyle inter-
Azhar Farooqi declare no support from any organisation for the submit- ventions to prevent or delay type 2 diabetes in people with impaired glucose toler-
ance: systematic review and meta-analysis. BMJ 334 (7588), 299.
ted work, no financial relationships with any organisations that might
Gray, L.J., Davies, M.J., Hiles, S., et al., 2012a. Detection of impaired glucose regulation and/
have an interest in the submitted work in the previous 3 years, and no or type 2 diabetes mellitus, using primary care electronic data, in a multiethnic UK
other relationships or activities that could appear to have influenced community setting. Diabetologia 55 (4), 959–966.
the submitted work. Melanie J Davies, Kamlesh Khunti and Thomas Gray, L.J., Khunti, K., Edwardson, C., et al., 2012b. Implementation of the automated
Leicester Practice Risk Score in two diabetes prevention trials provides a high yield
Yates declare no support from any organisation for the submitted of people with abnormal glucose tolerance. Diabetologia 55 (12), 3238–3244.
work and no financial relationships with any organisations that might Gray, L.J., Khunti, K., Williams, S., et al., 2012c. Let's Prevent Diabetes: study protocol for a
have an interest in the submitted work in the previous 3 years; MJD, cluster randomised controlled trial of an educational intervention in a multi-ethnic
UK population with screen detected impaired glucose regulation. Cardiovasc.
KK and TY were members (KK chair) of the NICE PH 38 (Preventing Diabetol.
type 2 diabetes: risk identification and interventions for individuals at Gusi, N., Olivares, P.R., Rajendram, R., 2010. The EQ-5D health-related quality of life ques-
high risk) Programme Development Group. tionnaire. In: Preedy, V.R., Watson, R.R. (Eds.), Handbook of Disease Burdens and
Quality of Life Measures New York. Springer, pp. 87–99.
Hex, N., Bartlett, C., Wright, D., Taylor, M., Varley, D., 2012. Estimating the current and fu-
ture costs of Type 1 and Type 2 diabetes in the UK, including direct health costs and
Transparency document indirect societal and productivity costs. Diabet. Med. 29 (7), 855–862.
Knowler, W.C., Barrett-Connor, E., Fowler, S.E., et al., 2002. Reduction in the inci-
dence of type 2 diabetes with lifestyle intervention or metformin. N. Engl.
The transparency document associated with this article can be J. Med. 346 (6), 393–403.
found, in the online version. Leal, J., Ahrabian, D., Gray, A.M., 2015. Cost Effectiveness of a Pragmatic Structured Educa-
tion Intervention for Type 2 Diabetes: Economic Evaluation of Data from the Let's
Prevent Trial Submitted to Preventative Medicine.
Lindström, J., Louheranta, A., Mannelin, M., et al., 2003. The Finnish Diabetes Prevention
Acknowledgments Study (DPS): lifestyle intervention and 3-year results on diet and physical activity. Di-
abetes Care 26 (12), 3230–3236.
The project was supported by the University of Leicester Clinical Morris, D.H., Webb, D., Achana, F., et al., 2013. Progression rates from HbA1c 6.0%–6.4%
and other prediabetes definitions to type 2 diabetes: a meta-analysis. Diabetologia
Trials Unit, the National Institute for Health Research Collaboration 56 (7), 1489–1493.
for Leadership in Applied Health Research and Care — East Midlands National Institute for Health and Clinical Excellence, 2003. Guidance on the use of Patient
(NIHR CLAHRC — EM), and the NIHR Leicester–Loughborough Diet, Education Models for Diabetes (Technology Appraisal 60). NICE, London.
National Institute of Health and Clinical Excellence, 2012. Preventing Type 2 Diabetes:
Lifestyle and Physical Activity Biomedical Research Unit, which is a Risk Identification and Interventions for Individuals at High Risk. NICE, London.
partnership between University Hospitals of Leicester NHS Trust, NHS, 2014. Five Year Forward View.
Loughborough University and the University of Leicester. Oldroyd, J.C., Unwin, N.C., White, M., Mathers, J.C., Alberti, K.G.M.M., 2006. Randomised
controlled trial evaluating lifestyle interventions in people with impaired glucose tol-
erance. Diabetes Res. Clin. Pract. 72 (2), 117–127.
Ramachandran, A., Snehalatha, C., Mary, S., et al., 2006. The Indian Diabetes Prevention
Appendix A. Supplementary data Programme shows that lifestyle modification and metformin prevent type 2 diabetes
in Asian Indian subjects with impaired glucose tolerance (IDPP-1). Diabetologia 49
(2), 289–297.
Supplementary data to this article can be found online at http://dx. Roe, L., Strong, C., Whiteside, C., Neil, A., Mant, D., 1994. Dietary intervention in primary
doi.org/10.1016/j.ypmed.2015.12.012. care: validity of the DINE method for diet assessment. Fam. Pract. 11, 375–381.
56 M.J. Davies et al. / Preventive Medicine 84 (2016) 48–56

Roper, N.A., Bilous, R.W., Kelly, W.F., Unwin, N.C., Connolly, V.M., 2001. Excess mortality in of multi-factorial cardiovascular risk intervention in people with type 2 diabetes
a population with diabetes and the impact of material deprivation: longitudinal, pop- mellitus detected by screening. Trials 11, 16.
ulation based study. BMJ 322 (7299), 1389–1393. Webb, D.R., Gray, L.J., Khunti, K., et al., 2011. Screening for diabetes using an oral glucose
Rubin, D.B., 2004. Multiple Imputation for Nonresponse in Surveys. John Wiley and Sons, tolerance test within a western multi-ethnic population identifies modifiable cardio-
New York. vascular risk: the ADDITION-Leicester study. Diabetologia 54 (9), 2237–2246.
Sintonen, H., Pekurinen, M., 1993. A fifteen-dimensional measure of health-related qual- World Health Organisation, 1999. Definition, Diagnosis, and Classification of Diabetes
ity of life (15D) and its applications. In: Walker, S.R., Rosser, R.M. (Eds.), Quality of Life Mellitus and its Complications. Report of a WHO consultation. Part 1: Diagnosis and
Assessment: Key Issues in the 1990s. Kluwer Academic Publishers, Dordrecht, Classification of Diabetes Mellitus. Report. World Health Organisation, Geneva.
pp. 185–195. World Health Organisation, 2011. Use of glycated haemoglobin (HbA1c) in the diagnosis
Tudor-Locke, C., Bassett, D.R., 2004. How many steps/day are enough? Preliminary pe- of diabetes mellitus (Geneva).
dometer indices for public health. Sports Med. 34, 1–8. Yates, T., Davies, M., Gorely, T., Bull, F., Khunti, K., 2009. Effectiveness of a pragmatic edu-
Tuomilehto, J., Lindstrom, J., Eriksson, J.G., et al., 2001. Prevention of type 2 diabetes cation programme aimed at promoting walking activity in individuals with impaired
mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N. glucose tolerance: a randomized controlled trial. Diabetes Care 32 (8), 1404–1410.
Engl. J. Med. 344 (18), 1343–1350. Zigmond, A.S., Snaith, R.P., 2006. The hospital anxiety and depression scale. Acta Psychiatr.
Webb, D.R., Khunti, K., Srinivasan, B., et al., 2010. Rationale and design of the ADDITION- Scand. 67, 361–370.
Leicester study, a systematic screening programme and randomised controlled trial