Open Access
Open Peer Review

This article has Open Peer Review reports available.

How does Open Peer Review work?

The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial

  • Tina Drud Due1Email author,
  • Thorkil Thorsen1,
  • Marius Brostrøm Kousgaard1,
  • Volkert Dirk Siersma1 and
  • Frans Boch Waldorff1
BMC Family Practice201415:65

DOI: 10.1186/1471-2296-15-65

Received: 7 January 2014

Accepted: 3 April 2014

Published: 9 April 2014

Abstract

Background

The Danish health care sector is reorganising based on disease management programmes designed to secure integrated and high quality chronic care across hospitals, general practitioners and municipalities. The disease management programmes assign a central role to general practice; and in the Capital Region of Denmark a facilitator-based intervention was undertaken to support the implementation of the programmes in general practice. The purpose of the study was to assess the effectiveness of this semi-tailored facilitator-based intervention.

Method

The study was a stepped-wedge, randomised, controlled trial among general practices in the Capital Region of Denmark. The intervention group was offered three one-hour visits by a facilitator. The intervention was semi-tailored to the perceived needs as defined by each general practice, and the practices could choose from a list of possible topics. The control group was a delayed intervention group. The primary outcome was change in the number of annual chronic disease check-ups. Secondary outcomes were: changes in the number of annual check-ups for type 2 diabetes (DM2) and chronic obstructive pulmonary disease (COPD); changes in the number of spirometry tests, changes in the use of ICPC diagnosis coding and patient stratification; sign-up for a software program for patient overview; and reduction in number of practices with few annual chronic disease check-ups.

Results

We randomised 189 general practices: 96 practices were allocated to the intervention group and 93 to the delayed intervention group. For the primary outcome, 94 and 89 practices were analysed. Almost every outcome improved from baseline to follow-up in both allocation groups. At follow-up there was no difference between allocation groups for the primary outcome (p = 0.1639). However, some secondary outcomes favoured the intervention: a higher reported use of ICPC diagnosis coding for DM2 and COPD (p = 0.0050, p = 0.0243 respectively), stratification for COPD (p = 0.0185) and a faster initial sign-up rate for the software program.

Conclusion

The mixed results from this study indicate that a semi-tailored facilitator-based intervention of relatively low intensity is unlikely to add substantially to the implementation of disease management programmes for DM2 and COPD in a context marked by important concurrent initiatives (including financial incentives and mandatory registry participation) aimed at moving all practices towards changes in chronic care.

Trial registration

ClinicalTrials.gov: NCT01297075

Keywords

Disease management programmes Facilitation Implementation RCT Outreach visits General practice Diabetes COPD

Background

About a third of the Danish population has at least one chronic disease and 70–80% of the resources in the Danish health care sector are used on these patients [1]. Disease management programmes (DMPs) based on the Chronic Care Model [2] have been developed in all regions of Denmark [3]. The DMPs outline a systematic, proactive approach to chronic care including a division of tasks between general practitioners, hospitals and municipalities. The programmes stress the need for population-based patient registration; annual chronic disease check-ups; and stratification of patients into three levels according to risk of complications, complexity, and state of the disease [4, 5].

The implementation of DMPs is complex and may be aided by facilitator-based interventions consisting of outreach visits in general practice by an external person providing information and support. Facilitator-based interventions have been used to support development and implementation of guidelines in general practice since the 1980s in England [6, 7] and have since been used in numerous interventions internationally and in Denmark [635]. Several studies have shown that facilitator-based interventions have contributed to changes in general practice [11, 19, 23, 26, 32, 36], while others have found no effect [20, 22, 28]. Although previous studies are very heterogeneous in terms of content and intensity of the interventions and the role of the facilitator, a new systematic review and meta-analysis of practice facilitation concludes that facilitation has a moderately robust effect on the adoption of guidelines [11]. A recent Cochrane review of educational outreach visits, a concept closely related to facilitation, concludes that the most consistent effects concern the prescribing of medicine, while the effects of facilitation are less clear and more varied in other areas such as prevention [25]. There are only few randomised controlled trials of interventions using practice facilitation as an implementation strategy to enhance the implementation of the chronic disease management [37, 38]. Thus, the aim of this study was to examine the effectiveness of a semi-tailored facilitator-based intervention developed by the Capital Region of Denmark to support the implementation in general practice of DMPs for chronic obstructive pulmonary disease (COPD) and type 2 diabetes (DM2).

Method

Setting

The Danish health care system is primarily tax financed and offers residents free-of-charge access to general practice (GP) and public hospital services. Almost all residents (98%) are registered with a GP, who serves as the primary care provider and a gatekeeper for patients’ referral to specialists and hospitals. The study population was all 762 general practices in the Capital Region of Denmark on 1 December 2010. We consecutively included practices that signed up for facilitation visits and completed a baseline questionnaire from January 2011 until we reached the 189 practices needed for the study. We excluded practices in which a facilitator worked and practices that had participated in a pilot study.

Study design

The study is a stepped-wedge, randomised, controlled trial (RCT) where the intervention is rolled out to groups sequentially and data are collected at baseline and before new groups receive the intervention. Groups that have not yet received the intervention thereby function as control groups [39]. The stepped-wedge design has been used in several studies in the health care sector and is perceived as relevant either for ethical or practical/logistical reasons [40].

The intervention was developed and implemented by the Region. The outcome measures were chosen in collaboration between the Region and the researchers. The design and execution of the RCT and the data analysis were performed independently by the researchers.

The practices were randomly allocated to facilitator visits in 2011 (intervention) or to facilitator visits in 2012 (ratio 1:1).

Ethics

The Danish Data Protection Agency and the Health Research Ethics Committee were informed about the study and found it unnecessary to report (j.nr. 2011-41-5958 and H-C-FSP-2011-04). The study is reported in ClinicalTrials.gov (NCT01297075). The Committee of Multi Practice Studies in General Practice in Denmark approved the study and recommended GPs’ to participate. (MPU 7–2011).

The intervention

The aim of the intervention was to support the implementation in general practice of the disease management programmes for COPD and DM2. The intervention focused on the GPs’ role as coordinator of care, patient stratification, a proactive approach, and a systematic organisation of the workflow and division of tasks in general practice concerning chronic disease check-ups. The intervention was semi-tailored to the perceived needs as defined by each general practice.

The intervention consisted of two phases: 1) the facilitator education and development of a toolbox and 2) the facilitator visits.

Facilitators

A total of 14 general practitioners (GPs) were recruited as facilitators as well as one organisational consultant. One of the 14 GPs had an additional education in organisational development. The facilitators were remunerated for participating in the development of the intervention and subsequently according to their number of visits.

Facilitator education and development of a toolbox

The training of the facilitators was multifaceted and consisted of an educational programme, workgroups, a pilot phase and on-going network meetings.

The educational programme consisted of about 40 hours of meetings (a weekend seminar and 10 three-hour meetings) and was composed of two major activities: a professional update in relation to the DMPs and related tools; and education on the act of facilitation including coaching, communication skills, meeting management and development processes.

Concurrently, the facilitators met in workgroups and developed various tools for use during the visits (e.g. PowerPoint slides introducing the different topics, descriptions of methods applicable to the facilitator, relevant brochures for hand-outs and practice management development tools). Each facilitator used 10 hours to develop the toolbox.

In a pilot phase, the facilitators practiced their skills and tools in 11 selected practices.

During the project period, three-hourly network meetings were held quarterly. The meetings aimed at further education, adjustments of the tools, and discussions among the facilitators concerning their role and experiences at the visits.

The facilitator visits

All GPs in the Region were informed about the intervention via postal letters, news mails, professional meetings and advertisements on the Region’s web pages. Each participating practice was offered up to three facilitator visits of each one hour. The visits were free of charge and the practice was compensated for lost income (approx. €200 per doctor in the practice per visit).

Before the visits

Before the allocation, the practices completed an online baseline questionnaire containing items about practice characteristics, annual check-ups for DM2 and COPD, diagnosis coding, patient stratification, and use of a software program for patient overview and quality data (Sentinel Data Capture). Close to the first visit, each practice completed a second online questionnaire that focused on their knowledge of the DMPs, division of tasks in the practice, collaboration with the municipalities, and suggestions for topics for the visits. A report containing the responses was generated and sent to the facilitator and the practice. The questionnaires provided the facilitators with an impression of the developmental state of the practice in relation to the DMPs, and this aided them in selecting potentially relevant elements from the toolbox. The responses could be brought into the discussions at the visits, but the questionnaires were not meant as an audit tool.

Prior to the first visit, the facilitator contacted the practice regarding the choice of topics to be discussed. The practice could choose one or more topics from a range of predefined topics within the scope of the DMPs for COPD and DM2:

  • Workflow and internal division of tasks in relation to regular and annual chronic disease check-ups

  • The software program for patient overview and quality data (Sentinel Data Capture) [41]

  • Diagnosis coding

  • Stratification of patients

  • Leadership and organisation

  • Collaboration with municipalities and hospitals

  • The role of GPs as coordinators of care

The visits

It was the intention that all doctors and staff members in the practice participated during the visits.

At the first visit, the facilitators were to assist in defining goals for practice development and in choosing suitable means to achieve them. During the subsequent meetings, the facilitators were to encourage and support a process of change. They were to act as discussion partners and colleagues rather than experts. They could also demonstrate relevant instruments for achieving the goals and recommend courses or other consultants for more specialised assistance.

After each visit the facilitator provided the practice with a standardised visit report containing the topics discussed and the goals and tasks agreed upon at the visit. It was a central element of the intervention that the practice was to work with the chosen tasks in the time until the next visit.

Outcome measures

Data were retrieved from the Danish National Health Service Register (DNHSR) [42], the Danish Quality Unit of General Practice [41] and from questionnaires. The DNHSR is used to manage the National Health Insurance covering the primary health care sector and, in particular, is used for settling accounts with providers.

The questionnaires were tested for relevance and applicability in a pilot study with 11 practices. The baseline questionnaires were collected before randomisation, and registry data were collected for each practice in the three months up to randomisation. The follow-up registry data were collected a year later, in the three months equivalent to the baseline period. The follow-up questionnaires were sent out on 15 February 2012. Thus, the follow-up was 12 months and the intervention 9 months. The delayed intervention group did not receive the intervention until after the follow-up registry data and questionnaires had been collected (Figure 1).
https://static-content.springer.com/image/art%3A10.1186%2F1471-2296-15-65/MediaObjects/12875_2014_Article_1180_Fig1_HTML.jpg
Figure 1

Timetable.

The primary outcome

  1. 1.

    Change in the number of annual chronic disease check-ups per 100 patients affiliated with the practice (DNHSR).

     

The annual chronic disease check-up is a structured consultation dealing with the status and progression of the disease and the disease-specific risk factors. The check-up also includes a dialogue about the patient’s wellbeing, illness experience, lifestyle and medical treatment. Lastly, it covers patient stratification and an assessment of the need for adjusting treatment and/or treatment goals. Stratification is a grouping of patients based upon their care needs’. It consists of an assessment of the current severity and complexity of the disease (including risk of complications and comorbidity) and an assessment of the patient’s capacity for self-care. The stratification is used to divide the responsibility for treatment between the general practitioner, the hospital and the municipality.

The annual chronic disease check-up is a core element in all DMPs [3], and we hypothesised that the facilitator visits would increase the use of annual chronic disease check-ups. Therefore, this was chosen as proxy measure of the implementation of the systematic approach to chronic care as emphasised in the programmes. We were able to retrieve this information directly from the DNHSR, independently of GPs’ assessments.

The secondary outcomes

  1. 2.

    Reduction in the number of practices with few (less than 1%) annual chronic disease check-ups per 100 patients affiliated with the practice (DNHSR).

     
  2. 3.

    Change in the number of spirometry tests per 100 patients affiliated with the practice (DNHSR).

     
  3. 4.

    Change in the number of annual check-ups for DM2 and COPD (self-reported).

     
  4. 5.

    Sign-up to the Sentinel Data Capture (a software program that automatically collects patient data from the GPs’ electronic health record system) [41] (register based).

     
  5. 6.

    Changes in the use of ICPC diagnosis coding for DM2 and COPD (self-reported).

     
  6. 7.

    Changes in the use of stratification of patients with DM2 and COPD (self-reported).

     

We hypothesised that the intervention would decrease the number of practices with few annual chronic disease check-ups (outcome measure 2), and increase the uptake of several tools indicative of a more systematic approach to chronic care (outcome measure: 3, 4, 5, 6 and 7), because these tools are core elements of the DMPs and were among the optional topics for the visits.

Sample size

The power calculation was based on data from the Danish National Health Service Register on the use of annual chronic disease check-ups during 1 January 2010–1 December 2010. Two parameters were calculated.
  1. 1.

    To increase the number of annual chronic disease check-ups from 1.25 to 2.0 consultations per 100 patients per quarter.

     
  2. 2.

    To reduce the number of practices with less than 1% annual chronic disease check-ups per 100 patients per quarter from 40% to 20%.

     

In accordance with the first parameter, there was a need for 128 practices at a power of 80% and a significance level of 5%. Using the second parameter, there was a need for 163 practices at the same power level. We estimated a dropout rate of 10% and therefore included 189 practices.

Randomisation and allocation concealment

Eligible practices were stratified by practice type (solo or group practice) and geographical location by using SAS version 9.2. The allocation of practices was done by an external organisation (Danish College of General Practitioners) independently of the research group. Due to the nature of the intervention it was not possible to conceal the intervention for the practices and the facilitators.

Statistical methods

Differences in the use of annual chronic disease check-ups and spirometry between allocation groups were assessed using t-tests. Differences between allocation groups regarding the rest of the outcome measures were assessed using chi-squared tests. All differences were assessed at baseline and at follow-up separately. The difference in sign-up rate to the Sentinel Data Capture over time was visualised in a Kaplan-Meier plot and analysed with a log-rank test. All statistical analyses were done using SAS version 9.2 (SAS Institute Inc., Cary, NC).

Results

Trial flow

Of the 762 practices in the Capital Region of Denmark, 189 were included in the RCT (Figure 2). There were no significant differences in practice characteristics or outcome measures between allocation groups at baseline (Table 1). Six practices dropped out of the study due to retirement and were excluded from the analysis and 12 practices did not answer the questionnaires and are represented only in the registry data. There were no differences in practice characteristics or baseline measures between the analysed practices and those lost to follow-up. In the delayed intervention group, 13 practices received visits during the intervention period, either because they were collaborating with practices in the intervention group or because they did not agree to delay their participation in the intervention.
https://static-content.springer.com/image/art%3A10.1186%2F1471-2296-15-65/MediaObjects/12875_2014_Article_1180_Fig2_HTML.jpg
Figure 2

Trial flow.

Table 1

Practice characteristics and the distribution of outcomes at baseline

 

Intervention group (N = 96)

Delayed intervention group (N = 93)

P-value

Practice characteristics (Regional registry data)

Type of practice

  

0.7605

   Solo

64%

67%

 

   Others

36%

33%

Ratio of doctors to patients (median (IQR))

1513 (1245-1768)

1503 (1078-1730)

0.1979

Proportion of practices with nurses

77%

70%

0.3227

Primary outcome measures (The Danish National Health Service Register)

   

Annual chronic disease check-ups per 100 patients affiliated with the practice N = 96/93 (median (IQR))

1.1 (0.4-2.4)

1.1 (0.3-2.3)

0.8372

Secondary outcome measures

Number of practices with less than 1% annual chronic disease check-ups N = 96/93 (The Danish National Health Service Register)

47%

47%

1.00

Spirometry test per 100 patients affiliated with the practice N = 96/93 (The Danish National Health Service Register) (median (IQR))

0.66 (0.2-1.1)

0.5 (0.1-1.0)

0.7156

Annual check-ups for Diabetes N = 9 6/93 (Questionnaires)

  

0.8185

   Yes – always

84%

87%

 

   Yes – sometimes

10%

8%

 

   No

5%

5%

 

Annual check-ups for COPD N = 96/93 (Questionnaires)

  

0.6900

   Yes – always

30%

30%

 

   Yes – sometimes

53%

48%

 

   No

17%

22%

 

Sign-up for the Sentinel Data Capture N = 96/93 (Regional registry data (DAK-E))

25%

22%

0.6085

ICPC diagnosis coding – Diabetes N = 96/93 (Questionnaires)

  

0.2391

   Yes – always

48%

40%

 

   Yes – sometimes

34%

32%

   No

17%

28%

ICPC diagnosis coding – COPD N = 96/93 (Questionnaires)

  

0.1311

   Yes – always

35%

26%

 

   Yes – sometimes

42%

39%

 

   No

23%

35%

 

Stratification – Diabetes N = 96/93 (Questionnaires)

  

0.3283

   Yes – always

14%

22%

 

   Yes – sometimes

36%

35%

 

   No

50%

43%

 

Stratification – COPD N = 96/93 (Questionnaires)

  

0.9324

   Yes – always

10%

11%

 

   Yes – sometimes

36%

33%

   No

53%

56%

 

The N at each outcome measure shows the number of practices in the analysis for the intervention/delayed intervention group.

Implementation of the intervention

Each practice could choose up to three one-hour visits. Seven practices divided the three hours into two visits; hence, the number of hours visited is used instead of number of visits. In the intervention group, 178 visits were undertaken during the intervention period. Of the 96 practices in the intervention group, 24% received all three hours of facilitation, 39% received two hours’, 29% received one hour and 9% had no visit. Some practices chose not to have more visits and others did not finish their visits during the intervention period but continued afterwards (46 of the 96 practices).

According to the visit reports made by the facilitators, several topics were usually discussed at the visits, but the number of topics typically decreased from the first to the third visit. The most common topic was the Sentinel Data Capture; other common topics were diagnosis coding, stratification, a website on municipal health services, and division of tasks.

Effectiveness of the intervention

With a few exceptions, all outcome measures improved from baseline to follow-up in both allocation groups (Tables 1 and 2).
Table 2

The distribution of outcomes at follow-up

 

Intervention group (2011) (N = 94)

Delayed intervention group (2012) (N = 89)

P-value

Primary outcome measures

   

Annual chronic disease check-ups per 100 patients affiliated with the practice N = 94/89 (median (IQR))

1.9 (0.9-3.9)

1.7 (0.7-3.5)

0.2788

Change in annual chronic disease check-ups per 100 patients affiliated with the practice N = 94/89 (median (IQR))

0.5 (0.0-1.9)

0.5 (0.0-1.3)

0.1639

Secondary outcome measures

   

Number of practices with less than 1% annual chronic disease check-ups N = 94/89

29%

33%

0.6314

Reduction in the number of practices with less than 1% annual chronic disease check-ups. N = 94/89

24%

18%

0.4403

Spirometry test per 100 patients affiliated with the practice N = 95/89 (median (IQR))

0.6 (0.2-1.2)

0.5 (0.1-0.8)

0.0835

Annual check-ups for Diabetes N = 89/82

  

0.2345

   Yes – always

92%

88%

 

   Yes – sometimes

8%

9%

 

   No

0%

4%

 

Annual check-ups for COPD N = 89/82

  

0.0787

   Yes – always

53%

37%

 

   Yes – sometimes

39%

49%

 

   No

8%

15%

 

Sign-up for the Sentinel Data Capture N = 94/89

71%

63%

0.2708

ICPC diagnosis coding – Diabetes N = 89/82

  

0.0050*

   Yes – always

87%

67%

 

   Yes – sometimes

12%

27%

 

   No

1%

6%

 

ICPC diagnosis coding – COPD N = 89/82

  

0.0243*

   Yes – always

73%

57%

 

   Yes – sometimes

25%

32%

 

   No

2%

11%

 

Stratification – Diabetes N = 89/82

  

0.0598

   Yes – always

27%

13%

 

   Yes – sometimes

43%

44%

 

   No

30%

43%

 

Stratification – COPD N = 89/82

  

0.0185*

   Yes – always

24%

11%

 

   Yes – sometimes

45%

39%

 

   No

31%

50%

 

The N at each outcome measure shows the number of practices in the analysis for the intervention/delayed intervention group.

*Indicate a significant difference between the allocations groups (P<0.05).

The use of annual chronic disease check-ups increased in both allocation groups (72% and 55% respectively), but there was no significant difference between the groups at follow-up (p = 0.2788) and no significant difference in the change from baseline to follow-up (p = 0.1639) (Table 2). There were no significant differences regarding the self-reported use of annual check-ups for DM2 and COPD (p = 0.2345; p = 0.0787 respectively), the reduction in the number of practices with few annual chronic disease check-ups (p = 0.4403), or in the use of spirometry (p = 0.0835).

The self-reported use of ICPC diagnosis coding was significantly in favour of the intervention group for both DM2 (p = 0.0050) and COPD (p = 0.0243). The difference in the proportion of practices that reported always using ICPC diagnosis coding at follow-up was 20 percentage points for DM2 and 16 percentage points for COPD.

The self-reported use of stratification was significantly different at follow-up for COPD (p = 0.0185) and there was a non-significant trend for DM2 (p = 0.0598) (Table 2). The difference between the allocation groups was, respectively, 13 percentage points for COPD and 14 percentage points for DM2 in the proportion of practices that reported always using stratification at follow-up.

During the study period, the intervention group showed a faster sign-up rate to the Sentinel Data Capture; however, this did not result in a significant difference at the end of the study (Figure 3).
https://static-content.springer.com/image/art%3A10.1186%2F1471-2296-15-65/MediaObjects/12875_2014_Article_1180_Fig3_HTML.jpg
Figure 3

Sign up for Sentinel Data Capture. *Indicate a significant difference between the groups.

Because some practices in the intervention group received less than two hours of visits and some practices in the delayed intervention group received visits during the intervention period, we conducted a per protocol analysis comparing the practices in the intervention group that received a minimum of two hours’ visits (59 practices) with the practices in the delayed intervention group that received no visits (75 practices). In contrast to the prior results, there was no significant difference in the use of ICPC diagnosis coding for COPD in the per protocol analysis (p = 0.1104). However, the self-reported use of annual check-ups for COPD (p = 0.0232) and stratification for DM2 (p = 0.0026) became significantly different between the allocations groups.

Discussion

In this large, stepped-wedge RCT of the effect of a semi-tailored facilitator intervention to support the implementation of disease management programmes for COPD and DM2 in general practice, we found an improvement in both allocation groups for nearly all outcome measures. However, when comparing the two allocation groups there was no additional effect of the intervention on the use of annual chronic disease check-ups, which was the primary outcome. Nevertheless, some of the self-reported secondary outcomes were in favour of the intervention, indicating a modest intervention effect.

In this study we used process-related rather than patient-related outcome measures. We did this because the objective of the study was to assess the effectiveness of the intervention as a tool to enhance the implementation of the disease management programmes.

The design of the intervention was characterised by a high degree of flexibility, for example, the latitude of the practices in selecting topics, the latitude of the facilitators in shaping their facilitation approach, and the absence of a detailed description for standardising the structure and content of the visits. This flexibility is likely to have generated wide variation across sites, thereby reducing the focus of the intervention and hence its impact on individual outcome measures. While the visit reports show that the practices chose to focus on similar topics, it is not possible to assess variations in how the topics were actually addressed or the time and level of detail spent on each topic. Conversely, the flexible design may have contributed to the positive impact of the intervention on some of the secondary outcomes by enhancing the motivation of the practices (by allowing them to choose the topics they found the most relevant), and by allowing the facilitators to tailor their approach to the particular situation of each practice.

That all outcome measures improved in both allocation groups from baseline to follow-up is probably influenced by chronic care and DMPs being highly profiled concepts in both the Capital Region of Denmark and the Danish health care system as a whole during the study period. The collective agreement between the Danish Regions and the GPs’ central organisation involves agreements and financial incentives for improving chronic care, and during the intervention period it became mandatory for all practices in Denmark to sign up for the Sentinel Data Capture before 31 March 2013. Further, several workshops and courses on Sentinel Data Capture, diagnosis coding, COPD and DM2 were provided in the Capital Region during the intervention period. These concomitant factors are highly likely to have influenced the study outcome. However, influences from the broader political context are inevitable when studying health services interventions in a real life setting, and hence this study underlines the importance of undertaking RCTs.

Strengths of the study

Given that all outcome measures, with a few exceptions, improved from baseline to follow-up in both allocation groups, the study demonstrates the value of a randomised controlled trial design that exposes changes attributable to the intervention. If a simple before-after measurement had been applied, the effect would have been overestimated profoundly. In our study the stepped-wedge design was chosen because the facilitator-based intervention was a public intervention offered to all practices in the Capital Region of Denmark; as the Region assumed that the intervention would be beneficial, they did not consider a non-intervention group as an option.

We included a large number of practices and used a primary outcome that could be obtained from administrative registers independently of the participating practices. The dropout rate was relatively small and less than anticipated, which enhanced the power of the study.

Limitations of the study

The practices included in this study were among the first ones to sign up for the intervention. Thus both allocation groups likely represent practices more interested in making changes, and this probably contributed to the improvements in both allocation groups.

The power calculations were based on an increase in annual chronic disease check-ups. The change in the intervention group equals the change used in the power calculations, but we did not take into account the large improvement in the delayed intervention group. This indicates that despite our study being large, it may be underpowered.

It is possible that the effect of the intervention was reduced by some practices in the intervention group not receiving the intended intervention and some in the delayed intervention group receiving visits during the intervention period. The per-protocol analysis revealed minor changes in the results regarding some of the secondary outcome measures. However, a per-protocol analysis does not adhere to the principles of an RCT since it imposes a risk of unknown bias. Moreover, when excluding practices not adhering to the protocol, the study no longer measures the effectiveness of the intervention in a real life setting but the effect on a more selected group of practices. Lastly, that the per protocol analysis did not change the results of the primary outcome suggests that even absolute fidelity to the allocated intervention would not have changed the results considerably.

To secure a high degree of reliability, we chose to use registry data for our primary outcome. The data in the DNHSR are gathered from the GPs’ electronic invoices to the Regional Health Administration. Due to this link with the reimbursement system, DNSHR data on the most common basic services are generally assumed to be complete, albeit no validity studies have been published [42]. Annual chronic disease check-ups are, however, not registered in the DNHSR according to the specific disease but only as a generic code. Thereby, our primary outcome refers to all annual chronic disease check-ups in the practice and not exclusively to DM2 and COPD. Nevertheless, the annual check-ups are an essential part of the DMPs, and we consider them a good indicator of whether the practice has implemented a more systematic approach to chronic care as described in the programmes. Further, we presume that an increase in the overall number of annual chronic disease check-ups also reflects an increase in annual check-ups for DM2 and COPD, since these are some of the major diseases for which the GPs provide annual chronic disease check-ups and are the two first diseases for which DMPs were developed. Additionally, data from the self-reported questionnaires concerning annual check-ups specifically for DM2 and COPD support the result of our primary outcome.

Several of the secondary outcome measures are self-reported and the significant results are among these. With the current data it is not possible to test the accuracy of the self-reported data information. Additionally, some parameters relevant in relation to the DMPs are not included in the study, for example, the content of the annual check-ups, and how the stratification is actually applied. Such data were not available in the study period.

The generalizability of the study

The study is an effectiveness study of an intervention in a real life setting and this adds to the external validity and the possibility of generalising the result to other regions in Denmark. However, it is more difficult to generalize the results to health systems that are markedly different from the Danish setting in which GPs are private entrepreneurs primarily financed through Public Health Insurance and where the collective agreements with the health authorities play a significant role in shaping the services provided.

Comparison with the literature

A recent systematic review [11] finds a positive effect of practice facilitation, whereas the recent Cochrane review of educational outreach visits finds small but consistent effects concerning the prescription of medicine but more mixed results for other areas such as prevention [25]. Individual studies also find an effect in some outcome measures but not in others [4345]. Our study also has mixed results showing a lack of effect on the primary outcome but significant effects on some of the secondary outcomes. Most of the RCT studies in the recent systematic review of practice facilitators are small compared with our study, and the outcomes are primarily assessed by questionnaires and patient record audits in contrast to data from administrative registers, which are more independent of potential interests in the intervention.

In our study, the intervention consisted of up to three visits within a year. It was an essential component in the intervention that the practices were to work with the planned topics between the visits, for example, by designing and implementing new procedures and implementing the Sentinel Data Capture. In a Danish primary care context this is considered to be an intensive intervention. Nevertheless, compared with other practice facilitation studies [20, 36, 37, 43, 45], the intervention in this study can be regarded as being of low intensity. A systematic review of practice facilitation in primary care found that the number of facilitator visits varies substantially between interventions, from a few meetings to several meetings a month throughout a whole year [11]. While some studies with few visits do report an effect [30, 46, 47], the review finds a positive relationship between the effect of the interventions and the intensity of visits, but not between the effect and duration of the intervention [11]. Thus, the limited effects of the present intervention may partly be ascribed to the intervention’s relatively low intensity.

In this intervention the facilitators were all general practitioners, apart from one. Hence they were colleagues and had an understanding of the organization and conditions in general practice, and they had experience of change processes from their own practices. This was considered by the project initiator as critical to ensure the participation of the practices. Compared with previous studies, it is unusual that the facilitators were GPs (although a few other studies from Denmark have used GPs as facilitators [15, 33]). Most often the facilitators are practice nurses or practice assistants [11]. In several studies, however, the profession of the facilitator is not mentioned [11]. Our study was not designed to assess the influence of the profession of the facilitators.

Although the concept of facilitation is widely used, it is not well defined. Many different terms are used to describe facilitator-based interventions (e.g. “facilitators”, “educational facilitators”, “educational outreach visits”, “outreach facilitation”, “practice facilitators” and “academic detailing”) [17] and often similar terms are used to describe interventions that appear to be different. Thus, the studies included in the recent systematic review vary on several parameters such as the number of visits, the professional background of the facilitator, the flexibility of the intervention, the role of the facilitator, and the outcome measures applied. Further, the descriptions of the interventions are often superficial, which makes it difficult to assess the actual interventions and the variations between them. This seriously impedes efforts to assess the general effectiveness of practice facilitation. Thus, in practice “facilitation” can be translated (i.e. interpreted and enacted) in many ways by the participating actors during the design and implementation of a facilitator-based intervention. The translation involves both a structural-logistic aspect (related to the number of visits, the duration of the intervention and the profession of the facilitators) and a more conceptual aspect related to the facilitator’s role and the type of interaction with the practices. In addition to the lack of conceptual clarity, only few studies, such as the ones by Hogg et al., Watkins et al. and Baskerville et al., have undertaken process evaluation or studied the implementation of practice facilitation [10, 18, 29]. It has been recommended to undertake more process evaluations to better understand the learning processes related to facilitator visits and the variations in results from different effect studies [25]. Further, even fewer studies have explored how the role of the facilitator is carried out in practice [15, 27].

In future studies it is therefore crucial that both the intended and the implemented translation of the concept are thoroughly described. Hence, we also collected a range of qualitative data alongside the RCT to explore how the concept of facilitation and the role of the facilitator were translated at all levels of the intervention from project developers to the actual visits, and how and why changes were occurring (or not occurring) during and after the visits (study not completed).

Conclusion

The mixed results from this study indicate that a semi-tailored facilitator-based intervention of relatively low intensity is unlikely to add substantially to the implementation of disease management programmes for DM2 and COPD in a context marked by important concurrent initiatives (including financial incentives and mandatory registry participation) aimed at moving all practices towards changes in chronic care. The study also points to the importance of conducting RCTs when evaluating practice facilitation in a changing professional context because of an otherwise profound risk of overestimating the effect of a new intervention.

Abbreviations

COPD: 

Chronic obstructive pulmonary disease

DAK-E: 

Danish Quality Unit of General Practice

DMPs: 

Disease management programmes

DM2: 

Type 2 diabetes

DNHSR: 

Danish National Health Service Register

GP: 

General practitioner

ICPC: 

International classification method for primary care encounters

IQR: 

Interquartile range

RCT: 

Randomised controlled trial.

Declarations

Acknowledgement

We thank Willy Karlslund, data manager at the Research Unit for General Practice, for the help he provided in the data management and data analysis. Further, we thank the reviewers for helpful and constructive comments. We would also like to thank the Danish Research Foundation for General Practice, the Health Insurance Foundation, and the Research Foundation for Primary Care in the Capital Region of Denmark for funding the study.

Authors’ Affiliations

(1)
The Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen

References

  1. Kronisk sygdom: Patient, sundhedsvæsen og samfund [Chronic disease. Patient, Health Service and Society]. 2005, Copenhagen: The Danish Health and Medicines AuthorityGoogle Scholar
  2. Improving Chronic Illness Care. 2013, http://www.improvingchroniccare.org,
  3. Forløbsprogrammer for kronisk sygdom - Generisk model og forløbsprogram for diabetes: [Disease management programmes - generic model and Disease management programmes for Type 2 Diabetes]. 2008, Copenhagen: The Danish Health and Medicines AuthorityGoogle Scholar
  4. Forløbsprogram for Type 2 Diabetes - Hospitaler, almen praksis og kommunerne i Region Hovedstaden: [Disease management programme For Type 2 Diabetes - Hospitals, general practice and municipalities in the Capital Region of Denmark]. 2009, Hilleroed, Capital Region of DenmarkGoogle Scholar
  5. Forløbsprogram for KOL - Hospitaler, almen praksis og kommunerne i Region Hovedstaden: [Disease management programme for COPD - Hospitals, general practice and municipalities in the Capital Region of Denmark]. 2009, Hilleroed, Capital Region of DenmarkGoogle Scholar
  6. Fullard E, Fowler G, Gray M: Facilitating prevention in primary care. Br Med J (Clin Res Ed). 1984, 289: 1585-1587. 10.1136/bmj.289.6458.1585.View ArticleGoogle Scholar
  7. Petrova M, Dale J, Munday D, Koistinen J, Agarwal S, Lall R: The role and impact of facilitators in primary care: findings from the implementation of the Gold Standards Framework for palliative care. Fam Pract. 2010, 27: 38-47. 10.1093/fampra/cmp066.View ArticlePubMedGoogle Scholar
  8. Kronikerkompasset evalueringsrapport: [The chronic care compass an evaluation report]. 2012, Aarhus: The Research Unit for General Practice in AarhusGoogle Scholar
  9. Atkins EM, Duffy MC, Bain DJ: The practice characterization model: the importance of organizational life cycles and targeted interventions in general medical practice. Int J Health Plann Manage. 2001, 16: 125-138. 10.1002/hpm.619.View ArticlePubMedGoogle Scholar
  10. Baskerville NB, Hogg W, Lemelin J: Process evaluation of a tailored multifaceted approach to changing family physician practice patterns improving preventive care. J Fam Pract. 2001, 50: W242-W249.PubMedGoogle Scholar
  11. Baskerville NB, Liddy C, Hogg W: Systematic review and meta-analysis of practice facilitation within primary care settings. Ann Fam Med. 2012, 10: 63-74. 10.1370/afm.1312.View ArticlePubMedPubMed CentralGoogle Scholar
  12. Bunniss S, Gray F, Kelly D: Collective learning, change and improvement in health care: trialling a facilitated learning initiative with general practice teams. J Eval Clin Pract. 2011, 18: 630-636.View ArticlePubMedGoogle Scholar
  13. Burtonwood AM, Hocking PJ, Elwyn G: Joining them up: the challenges of organisational change in the professional politic of general practice. J Interprof Care. 2001, 15: 383-393. 10.1080/13561820120080508.View ArticlePubMedGoogle Scholar
  14. Crotty M, Whitehead C, Rowett D, Halbert J, Weller D, Finucane P, Esterman A: An outreach intervention to implement evidence based practice in residential care: a randomized controlled trial [ISRCTN67855475]. BMC Health Serv Res. 2004, 4: 6-10.1186/1472-6963-4-6.View ArticlePubMedPubMed CentralGoogle Scholar
  15. Dalsgaard T, Kallerup H, Rosendal M: Outreach visits to improve dementia care in general practice: a qualitative study. Int J Qual Health Care. 2007, 19: 267-273. 10.1093/intqhc/mzm033.View ArticlePubMedGoogle Scholar
  16. Frich JC, Hoye S, Lindbaek M, Straand J: General practitioners and tutors' experiences with peer group academic detailing: a qualitative study. BMC Fam Pract. 2010, 11: 12-10.1186/1471-2296-11-12.View ArticlePubMedPubMed CentralGoogle Scholar
  17. Harvey G, Loftus-Hills A, Rycroft-Malone J, Titchen A, Kitson A, McCormack B, Seers K: Getting evidence into practice: the role and function of facilitation. J Adv Nurs. 2002, 37: 577-588. 10.1046/j.1365-2648.2002.02126.x.View ArticlePubMedGoogle Scholar
  18. Hogg W, Baskerville N, Nykiforuk C, Mallen D: Improved preventive care in family practices with outreach facilitation: understanding success and failure. J Health Serv Res Policy. 2002, 7: 195-201. 10.1258/135581902320432714.View ArticlePubMedGoogle Scholar
  19. Hogg W, Lemelin J, Moroz I, Soto E, Russell G: Improving prevention in primary care: Evaluating the sustainability of outreach facilitation. Can Fam Physician. 2008, 54: 712-720.PubMedPubMed CentralGoogle Scholar
  20. Hogg W, Lemelin J, Graham ID, Grimshaw J, Martin C, Moore L, Soto E, O'Rourke K: Improving prevention in primary care: evaluating the effectiveness of outreach facilitation. Fam Pract. 2008, 25: 40-48.View ArticlePubMedGoogle Scholar
  21. Janssens I, De MM, Habraken H, Soenen K, van DM, Christiaens T, Bogaert M: Barriers to academic detailers: a qualitative study in general practice. Eur J Gen Pract. 2005, 11: 59-63. 10.3109/13814780509178239.View ArticlePubMedGoogle Scholar
  22. McCowan C, Neville RG, Crombie IK, Clark RA, Warner FC: The facilitator effect: results from a four-year follow-up of children with asthma. Br J Gen Pract. 1997, 47: 156-160.PubMedPubMed CentralGoogle Scholar
  23. Nagykaldi Z, Mold JW, Aspy CB: Practice facilitators: a review of the literature. Fam Med. 2005, 37: 581-588.PubMedGoogle Scholar
  24. Nazareth I, Freemantle N, Duggan C, Mason J, Haines A: Evaluation of a complex intervention for changing professional behaviour: the Evidence Based Out Reach (EBOR) Trial. J Health Serv Res Policy. 2002, 7: 230-238. 10.1258/135581902320432769.View ArticlePubMedGoogle Scholar
  25. O'Brien MA, Oxman AD, Davis DA, Haynes RB, Freemantle N, Harvey EL: Educational outreach visits: effects on professional practice and health care outcomes. Cochrane Database Syst Rev. 2000, Art.no.: CD000409. doi:10.1002/14651858.CD000409.pub2Google Scholar
  26. Qureshi N, Allen J, Hapgood R: A systematic review of educational outreach visits for non-prescribing interventions in general practice. Eur J Gen Pract. 2002, 8: 31-36. 10.3109/13814780209160826.View ArticleGoogle Scholar
  27. Rhydderch M, Edwards A, Marshall M, Elwyn G, Grol R: Developing a facilitation model to promote organisational development in primary care practices. BMC Fam Pract. 2006, 7: 38-10.1186/1471-2296-7-38.View ArticlePubMedPubMed CentralGoogle Scholar
  28. Stevens SA, Cockburn J, Hirst S, Jolley D: An evaluation of educational outreach to general practitioners as part of a statewide cervical screening program. Am J Public Health. 1997, 87: 1177-1181. 10.2105/AJPH.87.7.1177.View ArticlePubMedPubMed CentralGoogle Scholar
  29. Watkins C, Timm A, Gooberman-Hill R, Harvey I, Haines A, Donovan J: Factors affecting feasibility and acceptability of a practice-based educational intervention to support evidence-based prescribing: a qualitative study. Fam Pract. 2004, 21: 661-669. 10.1093/fampra/cmh614.View ArticlePubMedGoogle Scholar
  30. Zwarenstein M, Bheekie A, Lombard C, Swingler G, Ehrlich R, Eccles M, Sladden M, Pather S, Grimshaw J, Oxman AD: Educational outreach to general practitioners reduces children's asthma symptoms: a cluster randomised controlled trial. Implement Sci. 2007, 2: 30-10.1186/1748-5908-2-30.View ArticlePubMedPubMed CentralGoogle Scholar
  31. Praksisbesøg i Region Syddanmark: [practice visits in Region Southern Denmark]. 2013, http://www.regionsyddanmark.dk/wm306156,Google Scholar
  32. Dalsgaard T, Rosendal M: [Motivation and barriers to the use of facilitator visits in general practice]. Ugeskr Laeger. 2008, 170: 731-735.PubMedGoogle Scholar
  33. Eriksson T, Siersma VD, Logstrup L, Buch MS, Elwyn G, Edwards A: Documenting organisational development in general practice using a group-based assessment method: the Maturity Matrix. Qual Saf Health Care. 2010, 19: e37-10.1136/qshc.2007.026187.View ArticlePubMedGoogle Scholar
  34. Kousgaard MB, Thorsen T: Positive experiences with a specialist as facilitator in general practice. Dan Med J. 2012, 59: A4443-PubMedGoogle Scholar
  35. Smidth M, Christensen MB, Olesen F, Vedsted P: Developing an active implementation model for a chronic disease management program. Int J Integr Care. 2013, 13: e020-View ArticlePubMedPubMed CentralGoogle Scholar
  36. Lemelin J, Hogg W, Baskerville N: Evidence to action: a tailored multifaceted approach to changing family physician practice patterns and improving preventive care. CMAJ. 2001, 164: 757-763.PubMedPubMed CentralGoogle Scholar
  37. Dickinson WP, Dickinson LM, Nutting PA, Emsermann CB, Tutt B, Crabtree BF, Fisher L, Harbrecht M, Gottsman A, West DR: Practice facilitation to improve diabetes care in primary care: a report from the EPIC randomized clinical trial. Ann Fam Med. 2014, 12: 8-16. 10.1370/afm.1591.View ArticlePubMedPubMed CentralGoogle Scholar
  38. Parchman ML, Noel PH, Culler SD, Lanham HJ, Leykum LK, Romero RL, Palmer RF: A randomized trial of practice facilitation to improve the delivery of chronic illness care in primary care: initial and sustained effects. Implement Sci. 2013, 8: 93-10.1186/1748-5908-8-93.View ArticlePubMedPubMed CentralGoogle Scholar
  39. Brown CA, Lilford RJ: The stepped wedge trial design: a systematic review. BMC Med Res Methodol. 2006, 6: 54-10.1186/1471-2288-6-54.View ArticlePubMedPubMed CentralGoogle Scholar
  40. Mdege ND, Man MS, Taylor nee Brown CA, Torgerson DJ: Systematic review of stepped wedge cluster randomized trials shows that design is particularly used to evaluate interventions during routine implementation. J Clin Epidemiol. 2011, 64: 936-948. 10.1016/j.jclinepi.2010.12.003.View ArticlePubMedGoogle Scholar
  41. Schroll H, Christensen RD, Thomsen JL, Andersen M, Friborg S, Sondergaard J: The danish model for improvement of diabetes care in general practice: impact of automated collection and feedback of patient data. Int J Family Med. 2012, 2012: 208123-View ArticlePubMedPubMed CentralGoogle Scholar
  42. Andersen JS, Olivarius NF, Krasnik A: The Danish National Health Service Register. Scand J Public Health. 2011, 39: 34-37. 10.1177/1403494810394718.View ArticlePubMedGoogle Scholar
  43. Frijling BD, Lobo CM, Hulscher ME, Akkermans RP, Braspenning JC, Prins A, van der Wouden JC, Grol RP: Multifaceted support to improve clinical decision making in diabetes care: a randomized controlled trial in general practice. Diabet Med. 2002, 19: 836-842. 10.1046/j.1464-5491.2002.00810.x.View ArticlePubMedGoogle Scholar
  44. Kinsinger LS, Harris R, Qaqish B, Strecher V, Kaluzny A: Using an office system intervention to increase breast cancer screening. J Gen Intern Med. 1998, 13: 507-514. 10.1046/j.1525-1497.1998.00160.x.View ArticlePubMedPubMed CentralGoogle Scholar
  45. Mold JW, Aspy CA, Nagykaldi Z: Implementation of evidence-based preventive services delivery processes in primary care: an Oklahoma Physicians Resource/Research Network (OKPRN) study. J Am Board Fam Med. 2008, 21: 334-344. 10.3122/jabfm.2008.04.080006.View ArticlePubMedGoogle Scholar
  46. Dietrich AJ FAU, O'Connor GT FAU, Keller AF, Carney PA FAU, Levy DF, Whaley FS: Cancer: improving early detection and prevention. A community practice randomised trial. BMJ. 1992, 304: 687-691. 10.1136/bmj.304.6828.687.View ArticlePubMedPubMed CentralGoogle Scholar
  47. Modell M, Wonke B, Anionwu E, Khan M, Tai SS, Lloyd M, Modell B: A multidisciplinary approach for improving services in primary care: randomised controlled trial of screening for haemoglobin disorders. BMJ. 1998, 317: 788-791. 10.1136/bmj.317.7161.788.View ArticlePubMedPubMed CentralGoogle Scholar
  48. Pre-publication history

    1. The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2296/15/65/prepub

Copyright

© Due et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. 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.