Type 2 diabetes is an increasingly prevalent chronic illness and an important cause of avoidable mortality. The prevalence in the North East Region of England is estimated at 5.6% (
http://www.yhpho.org.uk/resource/view.aspx?RID=81090) equating to approximately 96 patients per full time general practitioner. A review of quality of care studies (including diabetes) in UK primary care concluded that ‘in almost all studies the process of care did not reach the standards set out in national guidelines or set by the researchers themselves’ [
1]. A national audit of diabetes care showed that less than 50% of patients received all nine key aspects of diabetes care, with considerable variability across the country [
2]. However, care for people with diabetes has been improving: data from the UK’s voluntary incentive system for promoting high-quality care—the Quality and Outcomes Framework (QOF)—shows that quality scores across the UK are high, though many indicators are arguably relatively undemanding. However, for the most demanding indicators (
e.g., relating to tighter levels of HbA1c or blood pressure control) QOF performance is lower. Some of this variability likely reflects patient physiology or behaviour, but also reflects variable clinician management.
Recognition of quality gaps has led to increased interest in implementation research (the scientific study of methods to promote the systematic uptake of research findings into routine clinical practice) over the past 15 years [
3,
4]. The transfer of research findings into practice is often slow, leading to gaps in quality of care [
5]. Interventions to improve the quality of care that patients receive can be effective, but previous studies provided little theoretical or conceptual rationale for their choice of intervention [
6] and only limited descriptions of the interventions and contextual data [
7]. We have argued [
8] that a poor understanding of potential barriers and enablers to implementation limits the capacity to design appropriately targeted interventions. The challenge for implementation researchers is to develop and evaluate a theory-based approach to intervention design that will offer a generalisable framework for research and support the choice and development of interventions. The present cluster randomised controlled trial represents a response to this challenge, building upon recently completed development work identifying theoretical constructs that consistently predict healthcare professional behaviour and that are potentially modifiable in an intervention [
9,
10].
Evidence base for intervention development to identify theory: the improving quality in diabetes (iQuaD) study
Results from our recently completed study directly informs this protocol [
9,
11]. The ‘improving Quality in Diabetes’ (iQuaD) study was a national, theory-based study of the structural, organisational and individual factors associated with the performance of six healthcare professional behaviours involved in managing type 2 diabetes in primary care in the UK. We identified six clinical behaviours covering a range of clinical activities (prescribing, non-prescribing), some of which were challenging (
e.g., controlling blood pressure and HbA1c that was above target despite other drug treatment), and reflected recommended best practice as described by national guidelines at the time. Data were collected by telephone interview to a practice study contact, postal questionnaire to practice staff, postal questionnaire to patients, and from patient records. Ninety-nine practices completed a telephone interview and responded to baseline questionnaires assessing constructs from a range of theories. Response rates for all surveys and all professional groups were >75%. Scores on beliefs about the six target behaviours were generally consistent with good practice [
9], indicating that care gaps are likely a function of difficulties in regulating clinical behaviour in context, rather than knowledge deficits. We also showed that there is sometimes a discrepancy between what clinicians report providing and what patients report receiving [
12].
We assessed practice attributes and a wide range of individually reported measures at baseline; measured clinical outcomes over the ensuing 12 months, and administered a number of measures of clinical behaviour at baseline and at 12 months. A principal finding of iQuaD was that there continues to be variability in these clinical behaviours and most clinical behaviours that we investigated were not being optimally performed (see Table
1).
Table 1
Targeted clinical behaviours and evidence of performance from the iQuaD study, and associated NICE quality standards
1. Over the past 12 months provided general education about diabetes for patients with type 2 diabetes | Received by 73% of patients (via patient survey) | QS1 - ‘People with diabetes and/or their carers receive a structured educational programme that fulfils the nationally agreed criteria from the time of diagnosis, with annual review and access to ongoing education’ |
2. Over the past 12 months provided advice about weight management to patients with type 2 diabetes whose BMI is above a target of 30 kg/m2 even following previous management | Received by 51% of people whose BMI was above a target of 30 kg/m2, even following previous management (via patient survey) | QS2 - ‘People with diabetes receive personalised advice on nutrition and physical activity from an appropriately trained healthcare professional or as part of a structured educational programme’ |
3. Over the past 12 months provided advice about self-management for patients with type 2 diabetes | Received by 68% of patients (via patient survey) | QS3 - ‘People with diabetes participate in annual care planning which leads to documented agreed goals and an action plan’ |
4. Over the past 12 months prescribed additional therapy for the management of glycaemic control (HbA1c) in patients with type 2 diabetes whose HbA1c is higher than 64 mmol/mol (8.0%) despite maximum dosage of two oral hypoglycaemic drugs | Received by 59% of people whose HbA1c was higher than 8.0%, despite maximum dosage of two oral hypoglycaemic drugs (via practice-held prescribing data) | QS4 - ‘People with diabetes agree with their healthcare professional a documented personalised HbA1c target, usually between 48 mmol/mol and 58 mmol/mol (6.5% and 7.5%), and receive an ongoing review of treatment to minimise hypoglycaemia’ |
QS5 - ‘People with diabetes agree with their healthcare professional to start, review and stop medications to lower blood glucose, blood pressure and blood lipids in accordance with NICE guidance’ |
5. Over the past 12 months prescribe additional antihypertensive drugs for patients with type 2 diabetes whose blood pressure is 5 mmHg above target of 140 mmHg (systolic) or 80 mmHg (diastolic) even following previous management | Received by 40% whose blood pressure (BP) is above a target of 140 mm Hg for Systolic BP or 80 mm Hg for Diastolic BP, even following previous management (via practice-held prescribing data) | QS5 - ‘People with diabetes agree with their healthcare professional to start, review and stop medications to lower blood glucose, blood pressure and blood lipids in accordance with NICE guidance’ |
6. Over the past 12 months examined foot circulation and sensation in the feet of patients with type 2 diabetes | Received by 91% of patients (via patient survey) | QS10 - ‘People with diabetes at risk of foot ulceration receive regular review by a foot protection team in accordance with NICE guidance’ |
| | QS11 - ‘People with diabetes with a foot problem requiring urgent medical attention are referred to and treated by a multidisciplinary foot care team within 24 hours’ |
Our results showed that in primary care, individual healthcare professionals’ beliefs about these behaviours predicted their clinical behaviour while their perceptions about their workplace did not; multilevel models also showed that variability was primarily between individual clinicians within general practices rather than between the practices [
10]. Based on these findings, an intervention targeting individuals within organisations rather than targeting their organisation may be more likely to be effective.
The present study is directly informed by considerable theoretical development work accomplished within the iQuaD study wherein we theorised and tested a model of healthcare professional behaviour that combined existing theoretical explanations of behaviour [
13]. We initially tested ‘as originally theorised’ versions of multiple different individual [
10], organisational [
14] and stress theories [
15], to investigate to what extent constructs from multiple theories predict six different behaviours in the same sample of health professionals (enabling direct comparison of findings across multiple behaviours). The research in iQuaD was itself based on previous theory-based research conducted with different samples of healthcare professionals [
16‐
19]. In addition, we tested each theory using multilevel modelling to simultaneously account for individual and practice-level variability in theoretical constructs and clinical behaviour.
We showed particular constructs within particular behavioural theories to be consistently predictive of healthcare professional behaviour, including intention/proximal goals and self-efficacy (social cognitive theory [
20]), post-intentional factors (action and coping planning [
21]), and habit [
22] (learning theory), theories which have a broader evidence base within psychology [
23‐
25] and implementation science [
26]. We tested organisational theories, including organisational justice, organisational citizenship, and team climate, and found that while organisational justice factors predicted some clinical behaviours cross-sectionally [
14], none of the organisational-level theories consistently predicted clinical behaviours at 12 month follow-up [
27].
We also hypothesised
a priori how some of these theories might combine. We developed and operationalised a dual process model of health professional behaviour that simultaneously tested a sequential reflective process involving how intention to perform a clinical behaviour was mediated through post-intentional factors (action planning, coping planning) alongside a parallel impulsive process accounting for the degree to which clinicians behave automatically [
13]. We demonstrated that, for most clinical actions tested, a dual process approach contributed to understanding how reflective and impulsive factors relate to clinical behaviours.
A separate literature has also investigated the role of multiple goal pursuit, investigating how clinicians [
28,
29] and other populations [
30,
31] manage competing and facilitating goals and priorities in the time available. This research demonstrated that goal conflict and facilitation are readily identified and predictive of clinical and health-related behaviours and will further inform the present intervention.
To drive forward further improvements in quality of care, the UK National Institute for Health and Care Excellence (NICE) published 13 Quality Standards (QS) for Type 2 diabetes that cover a broader range of areas of diabetes care than current QOF indicators. Eight of these directly relate to primary care and to clinical behaviours assessed in the iQuaD study, which suggests areas in which current care falls short of achieving these indicators. Table
1 describes the QSs that map onto the clinical behaviours that we have previously investigated and which will be targeted for change by the intervention described in this protocol, as well as rates of performance of each behaviour and the sources of data we have used to assess these outcomes.
iQuaD provided an unprecedented opportunity to test and develop further theory that could inform the design of an intervention to improve diabetes quality of care. On the basis of this theoretical development, we will develop, pilot, deliver, and evaluate a theory-based behaviour change intervention targeting GPs and practice nurses to promote high-quality diabetes care consistent with NICE quality standards.