Skip to main content

Using the Theoretical Domains Framework (TDF) to understand adherence to multiple evidence-based indicators in primary care: a qualitative study

Abstract

Background

There are recognised gaps between evidence and practice in general practice, a setting posing particular implementation challenges. We earlier screened clinical guideline recommendations to derive a set of ‘high-impact’ indicators based upon criteria including potential for significant patient benefit, scope for improved practice and amenability to measurement using routinely collected data. Here, we explore health professionals’ perceived determinants of adherence to these indicators, examining the degree to which determinants were indicator-specific or potentially generalisable across indicators.

Methods

We interviewed 60 general practitioners, practice nurses and practice managers in West Yorkshire, the UK, about adherence to four indicators: avoidance of risky prescribing; treatment targets in type 2 diabetes; blood pressure targets in treated hypertension; and anticoagulation in atrial fibrillation. Interview questions drew upon the Theoretical Domains Framework (TDF). Data were analysed using framework analysis.

Results

Professional role and identity and environmental context and resources featured prominently across all indicators whilst the importance of other domains, for example, beliefs about consequences, social influences and knowledge varied across indicators. We identified five meta-themes representing more general organisational and contextual factors common to all indicators.

Conclusions

The TDF helped elicit a wide range of reported determinants of adherence to ‘high-impact’ indicators in primary care. It was more difficult to pinpoint which determinants, if targeted by an implementation strategy, would maximise change. The meta-themes broadly underline the need to align the design of interventions targeting general practices with higher level supports and broader contextual considerations. However, our findings suggest that it is feasible to develop interventions to promote the uptake of different evidence-based indicators which share common features whilst also including content-specific adaptations.

Peer Review reports

Background

International comparisons indicate that strong primary healthcare systems are associated with better population health, narrowing disparities and reduced costs [1]. Clinical research can make a major contribution to improving patient and population health in primary care settings [2] but only if findings are routinely incorporated into practice. Often, quoted figures suggest that around a third of patients do not receive care based on existing scientific evidence and about a quarter receive unnecessary or potentially harmful care [3, 4]. Figures of this kind continue to be reported in the literature. For example, there are substantial mismatches between evidence and recommended practice in the prescribing of lipid lowering drugs for the primary prevention of cardiovascular disease [5] and the self-reported care of long-term conditions in older people [6]. The most recent comprehensive overview of the quality of care delivered in the United Kingdom (UK), the NHS Atlas of Variation, illustrates large geographical variations in care and outcomes across several clinical areas, including diabetes, stroke and cancer [7].

Dissemination of the best practice, usually via clinical guidelines, is necessary but seldom sufficient by itself to ensure implementation [8]. The context of general practice in the UK presents particular implementation challenges—given limited practice organisational capacity, increasing complexity of care and the dispersed and independent nature of practices. Furthermore, in 2012, we identified 107 clinical guidelines relevant to general practice produced by the National Institute for Health and Care Excellence (NICE) [9]. This multiplicity of guidelines presents problems both for patients (e.g. those with multiple conditions) [10] and general practices responsible for their implementation.

There is an intuitive case for tailoring implementation strategies to identified needs and barriers [11, 12], even if there is still insufficient evidence on whether tailoring does enhance effectiveness [13]. Implementation studies generally target one condition or guideline (e.g. hypertension, back pain), but it is uncertain how findings for one guideline condition can be applied to another [14]. This suggests that implementation strategies need to be developed for each clinical guideline or even each guideline recommendation given that barriers to implementation can vary markedly between individual recommendations [15].

It is impracticable and inefficient to invent an implementation strategy for every guideline or recommendation. Yet, there are problems with adopting ‘one size fits all’ approaches to implementation. The quality landscape in UK primary care has been dominated by the Quality and Outcomes Framework (QOF) which attaches financial incentives to quality indicators [16]. It is uncertain whether the variable effects of QOF on the quality of care sufficiently outweigh its costs and unintended consequences (e.g. reduced continuity of care, therapeutic inertia) [1720]. Ideally, implementation strategies are required which can be adapted to a range of targeted problems and sustainably integrated into available primary care systems and resources [21].

Theoretical frameworks offer a common language with which to characterise contexts, targeted problems and interventions in generalisable terms and hence guide the adaptation of implementation strategies. A wide variety of theories from behavioural science, economics and social marketing are available to understand clinical behaviour [22]. The Theoretical Domains Framework (TDF) was specifically developed to identify determinants of professional behaviour change [23]. This framework includes 11 key determinants from 35 different theoretical models of behaviour and includes knowledge, skills, beliefs about consequences, beliefs about capabilities, social influences, emotion, motivation/goals, professional role/identity, memory and decision processes, environmental context and resources and action planning. An updated version also includes optimism and reinforcement and separates motivation (intention) and goals [24].

Many theories applied to implementation focus on one of the roles of individual cognitions, social context, or organisational characteristics. The TDF goes at least some way to integrating these different levels of influence, albeit focusing on the perceptions of those whose behaviour needs to change as the primary target for intervention. There also remains the potential to target features of the physical or social environment if these are identified as key determinants of behaviour. Moreover, subsequent work has identified and mapped behaviour change techniques onto the TDF to support the development of an approach which allows intervention developers to (i) identify the key determinants for a particular behaviour and (ii) propose a set of behaviour change techniques that address these [25, 26].

A growing number of interview studies report using the TDF as the basis for identifying determinants of guideline adherence, either through structuring the interview schedule to capture the 11 framework domains [27, 28] or for structuring the analytical process [29, 30]. However, questions still remain about the best way to identify the key domains that might be targeted for intervention. Indeed, the aforementioned studies have identified that most, if not all, of the domains represent barriers to behaviour change at some level. Moreover, within the implementation literature, few interventions target a single behaviour, mostly focusing on a recommendation or a set of recommendations. Data resulting from such interview studies is often a complex matrix of numerous behaviours, multiplied by up to 11 influencing factors and deciphering which domains should be targeted in the intervention is not straightforward. The application of the TDF to identify influences on behaviour is still in its infancy, and as such, there is a need for researchers to adopt a more critical approach to its use.

We used the TDF to explore health professionals’ perceived determinants of adherence to a set of primary care indicators derived from clinical guidelines. We examined which determinants were specific to indicators, thereby suggesting a need for indicator-specific tailoring of implementation strategies and which were shared across all indicators, thereby suggesting the potential of incorporating common elements into implementation strategies across different indicators. In considering shared determinants which potentially represent general influences on implementation within primary care, we also looked for meta-themes that emerged when synthesising data from multiple indicators.

Methods

Design and setting

We conducted semi-structured interviews with primary care professionals in West Yorkshire, UK.

Indicator selection

We had earlier screened NICE guidelines and associated quality standards to derive a set of ‘high-impact’ indicators based on burden of illness, potential for significant patient benefit from improved practice, likelihood of cost savings without patient harm and feasibility of measuring change using routinely collected data [9]. We selected eight of these indicators for further implementation work. This paper focuses on the four indicators that we subsequently took forward to targeting for intervention (Table 1):

Table 1 Indicators used in interview study
  • Avoidance of risky prescribing, especially for non-steroidal anti-inflammatory drugs (NSAIDs) [29]

  • Treatment targets in type 2 diabetes mellitus [31]

  • Blood pressure targets in treated hypertension [32]

  • Use of anticoagulation in atrial fibrillation [33]

The latter three of these were broadly aligned with QOF targets whilst generally pushing them further (e.g. aiming for tighter than incentivised blood pressure control) in line with guideline recommendations.

Sample

We intended to conduct 60 interviews with each interview covering one of the above four indicators. To gain a range of perspectives within practice teams, we aimed for a total sample comprising 30 general practitioners (GPs), 15 practice nurses and 15 practice managers.

Practices were approached on the basis of having contributed to an earlier part of the research programme. This earlier cross-sectional study involved 89 practices selected at random from across West Yorkshire and examined existing adherence to a larger set of clinical indicators (including the four covered in the present study). No effort was required of practices; adherence was assessed using remotely extracted, routinely collected data and practices simply had to consent to sharing of data.

Invitations to participate in the present study were sent to the 89 practices, and we then provided staff from interested practices with further information and contacted them to arrange interviews. We emphasised that the interviews would not be a formal test of knowledge and that we were exploring recognised problems with following recommended practice. We offered all interviewees £80 and a certificate confirming participation in the study in compensation for their time and obtained written informed consent prior to each interview. Recruitment ran from September 2013 until June 2014.

Interview procedure

Each interview covered two indicators, each including one of the four addressed in this paper. Participants initially completed a brief form to gather information on their age group, gender, current role and years’ experience in general practice. One of three researchers (GL, JH and EJI) conducted each interview. The topic guide drew on the TDF (see Appendix) and included two to three questions for each of the 11 TDF determinants [23]. The topic guide aimed to elicit knowledge and typical behaviours around each indicator as well as participants’ experiences of barriers to and enablers of following recommended practice. We did not ask participants to provide information about their level of compliance with the recommendation because we anticipated that this might exaggerate self-presentation bias and a focus on external influences on behaviour (e.g. environmental context, social influence) if compliance was low.

Data analysis

All interviews were audio-recorded and transcribed verbatim. We used NVivo software to facilitate analysis [32]. The same three researchers who conducted the interviews also undertook the analysis (GL, JH and EJI). We analysed interview data taking a Framework Analysis approach comprising familiarisation, identification of a framework, indexing, charting and mapping and interpretation [33]. The framework was developed through an iterative process which incorporated the study aims, the TDF and detailed reading of interview transcripts. This approach allowed for the inclusion of both a priori (e.g. TDF determinants) and emergent codes (e.g. specific patient factors).

As the researchers conducting the interviews were also responsible for the analysis, the initial familiarisation stage began during the interview process. Sets of completed interview transcripts were then allocated to each researcher to ensure that all researchers covered the range of indicators. As part of the familiarisation process, and to ensure that overarching themes were not missed during coding, researchers read through each transcript before coding and wrote a brief summary document outlining the key themes and findings within each transcript. Following agreement between the researchers, additional codes and categories identified in the indexing and familiarisation stages were added to the framework. Indexing in this context involved coding hard copies of the interview transcripts. In the early stages of this process, face-to-face meetings ensured agreement in coding. Ten percent of transcripts (n = 6) were coded independently by each researcher, and any disagreements were resolved through discussion.

For the initial TDF analysis assessing determinants for individual indicators, two researchers (GL and JH) examined the data coded within the TDF domains. Tables were produced to highlight key thematic content, barriers and enablers within each TDF domain. The researchers independently prioritised the primary TDF domains for each indicator and resolved disagreements by discussion.

For the analysis to identify and assess meta-themes across multiple indicators, the same two researchers (GL and JH) further interrogated the data, including the additional codes and categories generated to produce the analytical framework. This resulted in five meta-themes which incorporated data coded within the TDF, as well as data not captured by the TDF. The two researchers (GL and JH) finalised the meta-themes through discussion with RL.

Results

We conducted 60 face-to-face interviews as planned, with an approximate ratio of 2:1:1 between GPs, practice managers and nurses respectively from a total of 31 general practices (Table 2). Interviews typically lasted around 30 min per indicator. Most participants were female (70 %) and aged between 40 and 49 years (38 %; Table 3). The mean number of years’ experience in general practice was 14 (range 1 to 33).

Table 2 Allocation of interview topics
Table 3 Participant characteristics

We present findings, firstly, examining key TDF determinants for each indicator and, secondly, summarising meta-themes that emerged when synthesising data from multiple indicators.

Theoretical domain determinants by indicator

Table 4 presents a more descriptive account of all TDF domain content and specific barriers and enablers for each indicator.

Table 4 Key content relating to the Theoretical Domains Framework for each indicator

Risky prescribing

Compared to other staff, GPs appeared more knowledgeable about risky prescribing (knowledge). Awareness of drug interactions and patient histories were important. For example, possessing up to date knowledge was viewed as central to medication reviews. Differences between professional groups were highlighted (social and professional roles and identity); for example, GPs tended to believe that they had the autonomy to deviate from guidance whereas nurse prescribers described stringent adherence due to threats of litigation. There was an overriding sense that meeting patient needs was the main driver of prescribing practice rather than unquestioning adherence to indicators. Interviewees highlighted beliefs that adherence ensures quality of care, patient health and patient safety and also helps protect the reputation of the practice (beliefs about consequences). The potential long-term gains for the NHS (e.g. reduced hospital admissions) associated with adhering to these recommendations were perceived to far outweigh the immediate costs (e.g. increased consultation time, prescribing costs). Key barriers relating to environmental context and resources included lack of time (e.g. to keep up to date with relevant educational activities) and decision aids, as well as inadequacies of communication systems with secondary care (e.g. communication of prescription changes). Enablers included pharmacist support, prescribing leads and external support from clinical commissioning groups (CCGs; bodies comprising practice members responsible for commissioning services and assuring the quality of primary care). Current information technology systems, alerting prescribers to comorbidities for instance, were perceived as sometimes unsupportive of intuitive cognitive processes (memory, attention and decision processes). Familiarity with individual patient records was perceived as being central to whether or not medication was prescribed.

Treatment targets in type 2 diabetes

Many healthcare professionals felt there were clinicians in the practice that due to their expertise had more of a designated role to manage patients with diabetes (social and professional roles and identity). They described referring patients to the diabetic lead, particularly for patients taking multiple medications. Again, meeting patients’ needs rather than adhering to strict targets was a main driver of behaviour. In terms of knowledge, the indicator items were described as familiar and part of standard practice, although some professionals were less aware of target HbA1c levels. There was a perception that targets lead to pressure which may affect rapport with the patient during consultations and negative outcomes for the patient, such as the side effects of medication (beliefs about consequences). However, helping patients achieve target outcomes resulted in job satisfaction. Key enablers for environmental context and resources included information technology systems within the practice, training and education available and CCG support whilst barriers included low staffing levels and high competing workloads. Pressure from QOF and benchmarking (as social influences) were acknowledged as motivating target achievement.

Anticoagulation in atrial fibrillation

Interviewees considered that patients with atrial fibrillation often present acutely and hence anticoagulation is often initiated in secondary care (knowledge and social and professional roles and identity). The relatively infrequent presentation or detection of atrial fibrillation in primary care meant that staff often felt lacking in sufficient experience to initiate treatment, compounded by relative difficulties in recalling relevant guidance (memory, attention and decision processes). Interviewees felt that it was not always appropriate to adhere to recommendations for all patients when considering factors such as age or whether patients were taking multiple medications (beliefs about consequences). Barriers related to environmental context and resources included inadequate communication between primary and secondary care whilst having clear lines of responsibility within the practice were enabling. Behavioural regulation was supported by computer prompts, templates, audit and medication reviews, although the specificity and integration of prompts within computerised patient records could be improved.

Blood pressure targets in treated hypertension

Professional ethics and threat of litigation from under-treatment were perceived as enablers (social and professional roles and identity). However, there was a broad recognition of the need to tailor targets and treatment plans to individual patients. Although adhering to relevant guidance increased workload, such as consultation duration, interviewees perceived medium- and long-term benefits in doing so (beliefs about consequences). Barriers related to environmental context and resources included the limited availability of home blood pressure monitors whilst enablers included the availability of training, particularly opportunities to gain motivational interviewing skills. Practice team and local network meetings facilitated adherence whilst shared decision making with patients could operate in either direction (social influences).

Meta-themes spanning multiple indicators

We identified five meta-themes which potentially represent general influences on evidence-based practice: (i) perceived nature of the job and norms of practice; (ii) internal and external sources of support; (iii) communication pathways and interaction; (iv) meeting the needs of patients; and (v) perceptions of indicators. Tables 5, 6, 7, 8 and 9 present illustrative interview excerpts.

Table 5 Interview excerpts reflective of the theme ‘Perceived nature of the job and norms of practice’
Table 6 Interview excerpts reflective of the theme ‘Internal and external sources of support’
Table 7 Interview excerpts reflective of the theme ‘Communication pathways and interaction’
Table 8 Interview excerpts reflective of the theme ‘Meeting the needs of patients’
Table 9 Interview excerpts reflective of the theme ‘Perceptions of recommendations’

Perceived nature of the job and norms of practice

When discussing the indicators and associated clinical behaviours, healthcare professionals tended to view the workload and burden associated with adherence as accepted and embedded components of general practice. Whilst professionals sometimes felt that the indicators were imposed upon consultations and that there was a limit as to what was achievable within a typical 10-min medical consultation, they understood their utility in helping meet QOF targets and recognised standards of practice. They further recognised that implementation could improve outcomes and reduce healthcare costs in the longer term. Awareness of the indicators encouraged familiarity with required care processes and subsequent ingraining in everyday practice.

Although professionals described similar impacts of meeting the indicators, approaches to implementation differed between professional groups. Whilst GPs acted relatively autonomously and felt able to deviate from policies and procedures to tailor patient care, nurses preferred to follow policies and procedures, often justifying this approach by referring to risk and the threat of litigation. Some GPs felt that system prompts for implementing indicators disrupted consultations and sometimes directed their focus away from issues important to patients or the original reason why patients consulted. In contrast, many nurses said that they relied on templates and prompts to ensure that they were delivering appropriate care. These contrasting approaches to implementation partly reflect the more structured nature of nurse consultations, generally designed to achieve processes. However, GPs also indicated that they individually felt less pressure to achieve QOF targets than nurses did.

Internal and external sources of support

Healthcare professionals perceived both internal and external sources of support as critical to successful implementation. This often took the form of specialised support within the practice where specific practice staff had specialised knowledge or were established leads for a clinical area. External support was provided through access to colleagues in secondary care or network meetings with other practices. These sources of support provided trusted points of reference where professionals could seek the opinion of more knowledgeable colleagues and share and learn from others’ experience. Other supports assisted implementation by prompting memory and regulating clinical behaviour. These were provided at the practice level by regular practice meetings and the development and use of internally developed prompts and templates and at the wider organisational level via information technology and system infrastructure provided by the CCG and other bodies.

Communication pathways and interaction

Many healthcare professionals believed that effective interaction and information sharing were key to successful implementation of the indicators. These required channels and skills to facilitate communication at three levels: between professionals and patients; between colleagues in a practice; and between primary and secondary care. Effective communication also depended on the clarity of care pathways and respective professional roles. However, some professionals felt that there was scope for improving how communication systems could provide support.

Meeting the needs of patients

Healthcare professionals evidently considered it important to take a holistic view of the patient when making decisions, irrespective of whether this resulted in deviating from recommended practice. This individualisation of patient care appeared driven by a strong sense of professional ethos and beliefs that it truly reflected quality of care and improved patient outcomes. Professionals, particularly GPs, also acknowledged that patient priorities, preferences for treatment and social and financial circumstances all influenced their practice and hence achievement of indicators.

Whilst the latter factors were largely captured by the social influences of TDF domain, other patient factors outside of professional control influenced indicator achievement. These included patients’ own education and knowledge around conditions, varying adherence to treatment and failures to attend pre-arranged consultations and clinics. Such influences appeared particularly relevant for indicators focussed on outcomes and targets, i.e. diabetes and blood pressure control.

Perceptions of indicators

The content and structure of indicators and associated clinical practice recommendations represented another important influence not captured by the TDF. Whilst some recommendations regarded as relatively clear and simple to follow facilitated implementation, others were considered as unnecessarily complex, lacking in clarity, or too lengthy—hindering their application within a time pressured environment. There were also concerns about frequent revisions to recommendations and subsequent impacts on abilities to recall required procedures and processes. Some professionals also discussed how their perceived reliability of the source affected their opinion about the credibility of recommendations.

Discussion

We identified a wide range of factors which can determine adherence to ‘high-impact’ indicators in primary care. Those related to social and professional roles and identity and environmental context and resources were prominent themes across all indicators, whilst the importance of other domains, for example, beliefs about consequences, social influences and knowledge varied across recommendations. We further identified five more general meta-themes important to primary care professionals in the implementation of all the indicators. Taken together, our findings suggest that it is feasible to develop implementation strategies for different evidence-based indicators which share common features whilst also requiring content-specific adaptations.

Whilst the theoretical influences on adherence showed some consistency across the four indicators, there were important variations. For example, environmental context and resources featured in discussions of all of the indicators. However, the specific belief contents varied considerably, with poor communication between primary and secondary care being a problem for the prescribing of anticoagulation for patients with atrial fibrillation, whereas for the management of hypertension, constraints on resources, particularly the limited availability of blood pressure monitors was identified. Social and professional roles and identity was also important across all indicators. However, this may be explained in part by the study methods, as we interviewed practice professionals with varying roles in implementing the indicators. When this was expressed during the interview, such utterances were then coded as social and professional roles and identity. This was particularly found to be the case for practice managers. Other determinants also prominent in the conversations with staff included beliefs about consequences, social influences, knowledge and memory, attention and decision processes, the latter being particularly relevant for prescribing decisions. We also identified areas where the domains were less evident. These included motivation, beliefs about capabilities, skills and emotion. It is perhaps unsurprising that motivation, albeit extrinsic, was not identified as being particularly important. The QOF and the NICE guidelines offer both the evidence base and the incentives to support behaviour change, and therefore, there was rarely a question about the willingness or intention to adhere to the indicators.

Whilst it was possible to identify those factors that influenced adherence to the four indicators, it is more difficult to be confident about the extent to which targeting a particular domain is likely to bring about most change in adherence to an indicator. This suggests, as others have done (e.g. [34]) that additional research may be necessary to determine which particular barriers or enablers should be prioritised in implementation strategies. Although not reported here, we subsequently undertook stakeholder workshops in which we fed back the findings of the interviews to help us better understand the opportunities for implementation strategies.

We identified five meta-themes from a synthesis of data across all four indicators which broadly represent cultural, professional and system influences on evidence-based practice. Some of these might only be amenable to change at higher organisational levels (i.e. beyond the practice team), such as external sources of support and communication pathways or even further upstream in the development and dissemination of guidance, particularly perceptions of indicators [35]. Nevertheless, our findings underline the value of opportunities to share knowledge and expertise and support via local information technology systems for more efficient communication across care pathways.

Internal practice norms and ways of working appear critical to implementation, especially shared understanding of professional roles and mutual awareness of respective strengths and limitations. For example, practices delegated responsibilities for managing more complex management decisions related to titrating diabetes treatment or initiating anticoagulation. General practitioner clinical autonomy was important in considering the needs of patients with multiple morbidities, which often require trade-offs between the cumulative harms and benefits of treatments [36].

Our interviewees consistently indicated the central role of patients for certain indicators, especially where outcomes partly or largely depend on patient behaviour. Many of our interviewees recognised the role of consultation and counselling skills in enabling patient behaviour change. First, patients influence professionals’ decisions indirectly, sometimes via assumptions the latter make about the values and preferences of their patients. Second, the patient’s own behaviour was frequently referred to as a barrier to indicator achievement. For example, blood pressure control is more difficult to achieve if a patient drinks alcohol excessively or does not adhere to prescribed medication. Thus, the motivation and goals of both professionals and patients may need to be addressed simultaneously if outcomes are to be optimised [37]. Interventions which target both patients and professionals appear more likely to achieve glycaemic control in diabetes than those targeting either group in isolation [38].

Professionals often discussed general perceptions of guidelines and indicators. Their general attitude could be described as a predisposition to view adherence as an appropriate goal to strive for. Many participants, but GPs in particular, acknowledged that whilst for the population the value of guidelines was clear, for some patients, perhaps those with comorbidities or complex needs, adherence to recommendations was likely to result in poorer outcomes. This perceived inflexibility has been reported in other studies and reviews of guideline compliance [15, 39]. The knowledge that patients need to act to achieve some of the targets recommended in guidance and incentivised by QOF, together with the view that adhering to these targets might not always be in the best interest of a specific patient, is likely to influence compliance even if this is not commonly expressed when discussing motivation to achieve a specific indicator. Therefore, communications to professionals promoting adherence to evidence-based indicators need to explicitly acknowledge that 100 % compliance is rarely achievable (where patients’ behaviour contributes to achievement of the target) or optimal (when patient exceptions are accounted for).

Strengths and limitations

The interview schedule, structured around the TDF domains provided a useful prompt for discussions with participants about the factors that influenced the uptake of recommended practice. The interview appeared to have good face validity, with participants actively engaged in discussions. However, because the indicators related to a set of behaviours (Table 1) or were actually presented as goals to be achieved (e.g. blood pressure control in hypertension) responses to questions rarely related to the enacting of a specific behaviour, e.g. taking a patient’s blood pressure during a consultation. Given that it is unlikely to be cost-effective to develop complex interventions in primary care that focus on one discrete behaviour, we were interested in whether an interview based on the TDF could provide useful data for understanding influences on adherence to indicators that might inform subsequent intervention development.

Given the interview schedule included all TDF domains, it is unsurprising that participants talked about all of these influences on behaviour. Prioritising those influences for attention in the design of an intervention was more difficult, however. This difficulty may be a function of the tautological nature of the approach we adopted here in which both the interview schedule and the framework for analysis were structured around the TDF. In other words, we actively encouraged participants to talk about each domain and analysed the data by looking for evidence that each domain was referenced in the talk of participants. Whilst this may have the advantage of prompting people to think about influences that might not come to mind (e.g. emotion), it did make the prioritising of domains for intervention development difficult. Simply asking participants to talk about the factors that influence their behaviour may be a better technique for identifying key domains. Disclosure of beliefs may also be affected by rapport built during the course of the interview. This approach also raised a question about the coding of both barriers and enablers within a domain. Is it more valuable for the purposes of intervention development to know about what inhibits people from engaging in the behaviour or to know about the things that support people to adopt the behaviour? Both seem important and, in fact, knowing about enablers might support the identification of specific opportunities for intervention. However, if when discussing a domain participants largely focus on those factors that help adherence, this might suggest that there is little room for improvement. In other words, it may be useful to code utterances as either barriers or enablers and assess the relative prevalence of barriers to enablers as a useful way of prioritising domains for intervention. This approach may not be straightforward as participants often did not differentiate between barriers and enablers in their talk, making such coding difficult. For example, a participant would talk about the skills needed to do X but would not say whether they or others had these skills and to what extent.

The TDF approach is of course based on the assumption that explanations of behaviour can be verbalised, that most individuals have the insight to do this and that these explanations resemble the actual influences on behaviour. Accepting the interview findings as ‘the truth’ that is not subject to post hoc rationalisation, self-presentation bias and so forth would be naïve. Although we attempted to minimise these influences on responses, it is impossible to eradicate the tendency to focus on external influences when explaining our own behaviour (fundamental attribution error).

We recognise that the TDF is one framework amongst many which purport to explain behaviour [22]. We would expect many of the themes we identified to map onto other frameworks and theories, for example, the inner and outer settings of the Consolidated Framework for Implementation Research [40]. However, the TDF offers the advantages of drawing attention to potentially modifiable determinants of behaviours and providing a basis for linking determinants to behaviour change techniques [25, 26].

We acknowledge the significant role that patients have as both influencers of health professional behaviour and as actors in their own right. Both roles affect the extent to which indicator targets are achieved. As such, we may have identified further barriers and enablers had we also interviewed patients. Our findings suggest the potential value of interventions for selected indicators that target both patients and professionals.

Conclusions

An interview schedule based on the TDF elicited a wide range of reported determinants of adherence to ‘high-impact’ indicators in primary care. Certain domains featured prominently across all indicators whilst others were indicator-specific. We further identified five general meta-themes important to primary care professionals in the implementation of all indicators; these themes indicate the need to align the design of interventions targeting general practices with higher level supports and broader contextual considerations. Challenges remain in prioritising barriers and enablers to target within implementation strategies. However, our findings suggest that it is feasible to develop interventions to promote the uptake of different evidence-based indicators which share common features whilst also including content-specific adaptations.

References

  1. 1.

    Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q. 2005;83(3):457–502.

    Article  PubMed  PubMed Central  Google Scholar 

  2. 2.

    Mant D, Del Mar C, Glasziou P, Knotterus A, Wallace P, van Weel C. The state of primary care research. Lancet. 2004;364:1004–6.

    Article  PubMed  Google Scholar 

  3. 3.

    Schuster MA, McGlynn EA, Brook RH. How good is the quality of health care in the United States? Milbank Q. 1998;76(4):517–63. 509.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  4. 4.

    Grol R. Successes and failures in the implementation of evidence-based guidelines for clinical practice. Med Care. 2001;39(8 Suppl 2):Ii46–54.

    CAS  PubMed  Google Scholar 

  5. 5.

    Wu J, Zhu S, Yao GL, Mohammed MA, Marshall T. Patient factors influencing the prescribing of lipid lowering drugs for primary prevention of cardiovascular disease in UK general practice: a national retrospective cohort study. PLoS One. 2013;8(7):e67611.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  6. 6.

    Steel N, Hardcastle AC, Clark A, Mounce LTA, Bachmann MO, Richards SH, Henley WE, Campbell JL, Melzer D. Self-reported quality of care for older adults from 2004 to 2011: a cohort study. Age Ageing. 2014;43:716–20.

    Article  PubMed  PubMed Central  Google Scholar 

  7. 7.

    Right Care [2015], NHS atlas of variation in healthcare. Published September 2015 http://www.rightcare.nhs.uk/atlas/downloads/2909/RC_nhsAtlasFULL_MED_290915.pdf. Accessed 5th June 2016.

  8. 8.

    Grimshaw JM, Eccles MP, Lavis JN, Hill SJ, Squires JE. Knowledge translation of research findings. Implement Sci. 2012;7:50.

    Article  PubMed  PubMed Central  Google Scholar 

  9. 9.

    Rushforth B, Stokes T, Andrews E, Willis TA, McEachan R, Faulkner S, Foy R. Developing ‘high impact’ guideline-based quality indicators for UK primary care: a multi-stage consensus process. BMC Fam Pract. 2015;16(1):156.

    Article  PubMed  PubMed Central  Google Scholar 

  10. 10.

    Hughes LD, McMurdo ME, Guthrie B. Guidelines for people not for diseases: the challenges of applying UK clinical guidelines to people with multimorbidity. Age Ageing. 2013;42(1):62–9.

    Article  PubMed  Google Scholar 

  11. 11.

    Grol R. Beliefs and evidence in changing clinical practice. BMJ. 1997;315:418–21.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. 12.

    Shojania KG, Grimshaw JM. Evidence-based quality improvement: the state of the science. Health Aff. 2005;24:138–50.

    Article  Google Scholar 

  13. 13.

    Baker R, Camosso-Stefinovic J, Gillies C, Shaw EJ, Cheater F, Flottorp S, Robertson N. Tailored interventions to overcome identified barriers to change: effects on professional practice and health care outcomes. The Cochrane Library. 2010.

  14. 14.

    Foy R, Eccles M, Jamtvedt G, Young J, Grimshaw J, Baker R. What do we know about how to do audit and feedback? Pitfalls in applying evidence from a systematic review. BMC Health Serv Res. 2005;5:50.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  15. 15.

    Lugtenberg M, Zegers-van Schaick J, Westert G, Burgers JS. Why don’t physicians adhere to guideline recommendations in practice? An analysis of barriers among Dutch general practitioners. Implement Sci. 2009;4:54.

    Article  PubMed  PubMed Central  Google Scholar 

  16. 16.

    Roland M. Linking physicians’ pay to the quality of care—a major experiment in the United Kingdom. N Engl J Med. 2004;351:1448–54.

    CAS  Article  PubMed  Google Scholar 

  17. 17.

    Campbell S, Reeves D, Kontopantelis E, Sibbald B, Roland M. Effects of pay for performance on the quality of primary care in England. N Engl J Med. 2009;361:368–78.

    CAS  Article  PubMed  Google Scholar 

  18. 18.

    Guthrie B, Inkster M, Fahey T. Tackling therapeutic inertia: role of treatment data in quality indicators. BMJ. 2007;335:542–4.

    Article  PubMed  PubMed Central  Google Scholar 

  19. 19.

    Serumaga B, Ross-Degnan D, Avery A, Elliott R, Majumdar SR, Zhang F, Soumerai SB. Effect of pay for performance on the management and outcomes of hypertension in the United Kingdom: interrupted time series study. BMJ. 2011;342:d108.

    Article  PubMed  PubMed Central  Google Scholar 

  20. 20.

    McLintock K, Russell AM, Alderson SL, West R, House A, Westerman K, Foy R. The effects of financial incentives for case finding for depression in patients with diabetes and coronary heart disease: interrupted time series analysis. BMJ Open. 2014;4(8):e005178.

    Article  PubMed  PubMed Central  Google Scholar 

  21. 21.

    Dawda P, Jenkins R, Varnam R. Quality improvement in general practice. London: Kings Fund; 2010.

    Google Scholar 

  22. 22.

    Grol RP, Bosch MC, Hulscher ME, Eccles MP, Wensing M. Planning and studying improvement in patient care: the use of theoretical perspectives. Milbank Q. 2007;85(1):93–138.

    Article  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Michie S, Johnson A, Abraham C, Lawton R, Parker D, Walker A. Making psychological theory useful for implementing evidence based practice: a consensus approach. Qual Saf Health Care. 2005;14:26–33.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  24. 24.

    Cane J, O'Connor D, Michie S. Validation of the theoretical domains framework for use in behaviour change and implementation research. Implement Sci. 2012;7:37.

    Article  PubMed  PubMed Central  Google Scholar 

  25. 25.

    Michie S, Johnston M, Francis JJ, Hardeman W, Eccles MP. From theory to intervention: mapping theoretically derived behavioural determinants to behaviour change techniques. Appl Psychol. 2008;57:660–80.

    Article  Google Scholar 

  26. 26.

    Michie S, Richardson M, Johnston M, Abraham C, Francis J, Hardeman W, Eccles MP, Cane J, Wood CE. The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions. Ann Behav Med. 2013;46(1):81–95.

    Article  PubMed  Google Scholar 

  27. 27.

    Patey AM, Islam R, Francis JJ, Bryson GL, Grimshaw JM. Anesthesiologists’ and surgeons’ perceptions about routine pre-operative testing in low-risk patients: application of the Theoretical Domains Framework (TDF) to identify factors that influence physicians’ decisions to order pre-operative tests. Implement Sci. 2012;7:52.

    Article  PubMed  PubMed Central  Google Scholar 

  28. 28.

    Murphy K, O'Connor DA, Browning CJ, French SD, Michie S, Francis JJ, Russell GM, Workman B, Flicker L, Eccles MP, et al. Understanding diagnosis and management of dementia and guideline implementation in general practice: a qualitative study using the theoretical domains framework. Implement Sci. 2014;9:31.

    Article  PubMed  PubMed Central  Google Scholar 

  29. 29.

    McSherry LA, Dombrowski SU, Francis JJ, Murphy J, Martin CM, O'Leary JJ, Sharp L. ‘It’s a can of worms’: understanding primary care practitioners’ behaviours in relation to HPV using the Theoretical Domains Framework. Implement Sci. 2012;7:73.

    Article  PubMed  PubMed Central  Google Scholar 

  30. 30.

    Islam R, Tinmouth AT, Francis JJ, Brehaut JC, Born J, Stockton C, Stanworth SJ, Eccles MP, Cuthbertson BH, Hyde C, et al. A cross-country comparison of intensive care physicians’ beliefs about their transfusion behaviour: a qualitative study using the Theoretical Domains Framework. Implement Sci. 2012;7:93.

    Article  PubMed  PubMed Central  Google Scholar 

  31. 31.

    National Institute for Health and Care Excellence. Type 2 diabetes: the management of type 2 diabetes. NICE Clin Guidel. 2009;87.

  32. 32.

    Krause T, Lovibond K, Caulfield M, McCormack T, Williams B. Management of hypertension: summary of NICE guidance. BMJ. 2011;343:d4891.

    Article  PubMed  Google Scholar 

  33. 33.

    Jones C, Pollit V, Fitzmaurice D, Cowan C. The management of atrial fibrillation: summary of updated NICE guidance. BMJ. 2014;348:g3655.

    Article  PubMed  Google Scholar 

  34. 34.

    Mazza D, Chapman A, Michie S. Barriers to the implementation of preconception care guidelines as perceived by general practitioners: a qualitative study. BMC Health Serv Res. 2013;13:36.

    Article  PubMed  PubMed Central  Google Scholar 

  35. 35.

    Michie S, Johnston M. Changing clinical behaviour by making guidelines specific. BMJ. 2004;328:343–5.

    Article  PubMed  PubMed Central  Google Scholar 

  36. 36.

    Wallace E, Salisbury C, Guthrie B, Lewis C, Fahey T, Smith SM. Managing patients with multimorbidity in primary care. BMJ. 2015;350:h176.

    Article  PubMed  Google Scholar 

  37. 37.

    Campbell NC, Murray E, Darbyshire J, Emery J, Farmer A, Griffiths F, Guthrie B, Lester H, Wilson P, Kinmonth AL. Designing and evaluating complex interventions to improve health care. BMJ. 2007;334(7591):455–9.

    Article  PubMed  PubMed Central  Google Scholar 

  38. 38.

    Shojania KG, Ranji SR, McDonald KM, Grimshaw JM, Sundaram V, Rushakoff RJ, Owens DK. Effects of quality improvement strategies for type 2 diabetes on glycemic control: a meta-regression analysis. JAMA. 2006;296(4):427–40.

    CAS  Article  PubMed  Google Scholar 

  39. 39.

    Baiardini I, Braido F, Bonini M, Compalati E, Canonica GW. Why do doctors and patients not follow guidelines? Curr Opin Allergy Clin Immunol. 2009;9(3):228–33.

    Article  PubMed  Google Scholar 

  40. 40.

    Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009;4:50.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

This paper summarises independent research funded by the National Institute for Health Research under its Programme Grants for Applied Research scheme (‘Action to Support Practices Implementing Research Evidence;’ RP-PG-1209-10040). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.

The ASPIRE programme team comprises Vicky Ward, Robert M. West, Amanda Farrin, Martin Rathfelder, Susan Clamp, Claire Hulme, Paul Carder, Judith Richardson, Tim Stokes, Ian Watt and Suzanne Hartley, in addition to the named authors.

Authors’ contributions

Rebecca Lawton and Robbie Foy developed the study protocol. This was revised with input from Rosie McEachan, Thomas Willis and Liz Glidewell. Gemma Louch, Jane Heyhoe and Emma Ingleson collected and analysed the data, with support from Rebecca Lawton and Thomas Willis. Rebecca Lawton wrote the first draft of the paper, with input from Gemma Louch and Jane Heyhoe. This was edited by all other authors. Robbie Foy wrote the second draft of the paper. All authors read and approved the final manuscript.

Competing interests

Robbie Foy and Liz Glidewell are editors of Implementation Science. All decisions about this manuscript were made by another editor.

Ethics approval and consent to participate

The study was favourably reviewed by Leeds Central Research Ethics Committee (REC ref: 12/YH/0254).

Author information

Affiliations

Authors

Consortia

Corresponding author

Correspondence to Rebecca Lawton.

Appendix

Table 10 Interview topic guide

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lawton, R., Heyhoe, J., Louch, G. et al. Using the Theoretical Domains Framework (TDF) to understand adherence to multiple evidence-based indicators in primary care: a qualitative study. Implementation Sci 11, 113 (2015). https://doi.org/10.1186/s13012-016-0479-2

Download citation

Keywords

  • Primary care
  • Diabetes
  • Hypertension
  • Prescribing
  • Atrial fibrillation
  • Theoretical Domains Framework
  • Guideline implementation
  • Qualitative
  • Interviews