A `combined framework’ approach to developing a patient decision aid: the PANDAs model

Shared decision making is an important component of patient-centred care [1]. It involves the exchange of information and negotiation between the patients and healthcare professionals to reach a consensus on a particular decision [2]. The decision making process is complex and different decision support interventions have been developed to facilitate this process [3].

Patient decision aids (PtDAs) are `interventions designed to help people make specific and deliberative choices among options (including the status quo) by providing information on the options and outcomes relevant to a person’s health status and implicit methods to clarify values’ [4]. They are different from patient health education materials which provide general health information about specific medical conditions including diagnosis, investigation and treatment. There are currently more than 500 PtDAs that have been developed worldwide, mainly in North America and Europe. [Decision Aids Library Inventory http://​decisionaid.​ohri.​ca].

The effectiveness and utility of a PtDA is dependent on its quality. The International Patient Decision Aid Standards (IPDAS) collaboration has identified appropriate quality indicators using a Delphi consensus method and this provides guidance on the assessment of the content, format and evaluation of the PtDA (Table 1) [5]. However, few PtDAs developers describe the development process per se and, of those which do, there is a wide variation in how PtDAs are actually developed. In addition, PtDAs are often developed without users’ involvement [6].

Although the IPDAS guideline provides a framework for the development of a PtDA, it does not provide guidance on which research methods to use; neither does it explicitly link the development of a PtDA to its evaluation and implementation. It is, therefore, necessary to combine these guidelines with other frameworks such as that produced by the UK Medical Research Council (UKMRC) framework which can link the development of the PtDA to its evaluation and implementation (Figure 1) [7],[8]. It is designed to guide the development and evaluation of complex interventions, such as PTDAs, and it stresses the important role of evidence, theories and modelling in the process of developing a complex intervention. Specific research methods are recommended to operationalise each step and it was for this reason that UKMRC Framework was used to complement the IPDAS guideline to develop and evaluate the PDA described in this study. In addition, the UKMRC framework highlights the importance of having a range of options before deciding which intervention to use and that it should be based on existing best available evidence. Before developing a decision support tool, it is important to reflect whether PtDA is the best tool to use to support decision making.

This paper described how the Patient and insulin Decision Aids (PANDAs) PtDA was developed for people with Type 2 Diabetes Mellitus [T2DM] based on the IPDAS criteria and UKMRC framework. The PANDAs study was conducted from 2007-2010 to develop the PtDA and to evaluate its effectiveness in the UK general practice [9].

We used a systematic and iterative approach to develop the PtDA by incorporating the IPDAS and UKMRC framework (Table 2). The UKMRC framework informed the methods used for the development of the PtDA, which is not stated in the IPDAS guideline. This section described in detail how the methods proposed by the UKMRC framework were used to operationalise each step: expert consensus (e.g. drafting of the PtDA), literature review (e.g. collating clinical evidence), individual interviews (e.g. needs assessment) and questionnaire surveys (e.g. acceptability and feasibility) (Table 2). This part of the PANDAs study was conducted in the general practice setting, Sheffield, England in 2007-2008. We obtained ethics approval from the National Health Service Research Ethics Committee (REC reference: 07/Q2308/53).

The members of the expert panel were selected based on their experience as patients or involvement in the care for people with type 2 diabetes in general practice (Table 2). The expert panel was responsible for reviewing the first and the subsequently revised drafts of the PtDA produced by the research team. This expert panel assessed the PtDA based on: (1) member’s experience as patients who were making or who had made the decision about starting insulin treatment; (2) healthcare professionals who had helped patients in making such decisions; and (3) key opinion leaders/policy makers who were involved in implementing health services for diabetes in General Practice. Members of the review panel comprised two patients, three healthcare professionals, three key opinion leaders and three representatives from the local Primary Care Trust and research network.

We conducted the literature review for two reasons; firstly, we aimed to identify the effective decision support interventions tools that were available; secondly, we wanted to search for latest clinical evidence on the pros and cons of the different treatment options to be incorporated into the PtDA. Therefore, two systematic review were conducted. As far as possible, we searched for systematic reviews and national or local clinical practice guidelines. In the event that they were not available, findings from individual studies were appraised, synthesised and summarised. The quality of the information was graded according to the "grading for evidence-based rheumatology" guidelines [13].

A qualitative methodology using individual in-depth interviews and focus group discussion was used for the needs assessment, acceptability and feasibility studies. A mixed methodology was used in this process to triangulate the data so that the PtDA could cater for the comprehensive needs of the users. Patients with type 2 diabetes who were at the point of deciding whether or not to start insulin therapy were recruited into the study. Healthcare professionals, including general practitioners, practice nurses and dieticians, who were involved in managing patients with type 2 diabetes, were also included in the study. For the needs assessment, the users were interviewed individually for their needs when making (or facilitating) a decision on insulin therapy. As for the feedback on the acceptability and feasibility of using the PtDA, both the patients and clinicians were interviewed after they had used the PtDA at the consultation.

One of the researchers (CJN) conducted all the interviews using a semi-structured interview guides, which were developed separately for patients and clinicians based on the literature review and expert opinion (Additional files 1 and 2). The interviews were audio-recorded, transcribed verbatim, checked and analysed using a thematic approach. NVivo 7 software was used to manage the qualitative data.

In addition, a bespoke questionnaire was used to capture quantitative feedback from the users on the acceptability and feasibility of using the PtDA. The users provided general and specific feedback on the content, layout and usability of the PtDA using a five-point Likert scale.

The development of the PANDAs PtDA demonstrates that it is feasible to use a combination of the IPDAS and the UKMRC frameworks to develop a PtDA based on users’ needs, evidence and theories (ODSF). The IPDAS recommendations provided evidence-based guidance on the development of the content and structure of the PtDA; the users’ healthcare decisional, social and emotional needs informed the selection of the content; and the ODSF provided a theoretical basis to link the development of the PtDA to the outcome measures for evaluation. Finally, the UKMRC framework for developing and evaluating complex interventions in Primary Care provided an overarching framework to guide the iterative process, inform the choice of research methods and ensure that the PtDA could be successfully developed, evaluated and implemented in practice.

The recent IPDAS collaboration evidence document reviewed the PtDAs assessed in the latest Cochrane Review [4] and found that only half of the PtDAs underwent field testing with patients and clinicians in the `real world’ and only 17% documented the method of reviewing and synthesising clinical evidence [18]. It also highlighted the fact that few PtDAs provide sufficient details to assess whether or not they comply with the IPDAS criteria.

This study describes systematically how the IPDAS criteria guided the development of the PANDAs PtDA. The use of a theoretical framework to develop PtDAs was not part of the IPDAS standards [6]. However in the recent update on the IPDAS evidence document, the use of a theoretical framework to develop PtDAs has been proposed [19]. Similarly, the UKMRC guidance on how to develop complex interventions such as PtDAs also recommends the use of theories, which help to describe, explain and predict how the PtDA might work. With the growing number of clinical trials on PtDAs, a comparison between PtDAs which have or do not have a theoretical framework, is possible and the revised IPDAS standards on the evaluation of PtDA quality can be used to address this task [20].

The decision on `what’ and `how much’ to include in a PtDA is also a challenging one. Conventionally, PtDAs have focussed on delivering accurate and balanced information about the treatment options to patients ie emphasising the information needs [21]. However, increasingly, there is growing evidence that people do not make decisions based only on the information they receive [22]. Other considerations such as the availability of healthcare, emotional and social support may also affect how patients make decisions and it is, therefore, crucial that these needs of patients are attended to during the decision making process. In this study, the findings from the needs assessment reinforces this.

When identifying the best evidence to be included in the PtDA, we encountered a few difficulties. Firstly, there was a lack of systematic reviews or trials comparing insulin with `make no change’ or `lifestyle modification’. This is because it would be unethical and unacceptable to compare insulin therapy with no treatment in people with uncontrolled diabetes. There is also a lack of trials comparing insulin with lifestyle changes as most of these trials are funded by the pharmaceutical industry which have a vested interest in promoting drug usage [23].

In addition, there are also substantial variations in the reported outcomes of studies. For example, some studies report hypoglycaemic episodes as the number of episodes per 100 person year, whilst others report outcomes as "proportions of people experiencing hypoglycaemic episodes over a year". In addition, few clinical trials report on outcomes that are important to patients, such as `symptom relief’, which are relevant to patients when making any sort of medical decision [24].

In the PANDAs PtDA a list of diabetic symptoms was included and patients were asked to indicate whether or not they had experienced any of the symptoms. Although the evidence was, to some extent, lacking, "diabetic symptom relief" was included the in the PtDA as this is advice that is routinely given by healthcare professionals and it has a plausible scientific basis. Future clinical trials should routinely report outcomes that are clinically important to patients.

In the PANDAs PtDA, the risks and benefits were communicated to patients using the accepted principles of risk communication such as using textual, number and visual display to present risks; using natural frequencies rather than percentages; and emphasising the interaction between patients and healthcare professionals [25]. There is an increasing recognition that people use both analytic and intuitive mechanisms to make decisions; so how people perceive risk may depend as much on logic and systematic analysis as well as emotions and heuristics [26].

Concerns were also raised about the purpose of the PtDA and whether it would be used to "convince" patients to start insulin treatment. Thompson PB has raised the same issue that risk communication might be used to encourage changes in peoples’ behaviour for the sake of public health gain [27]. However, the aim of the PANDAs PtDA was to help patients make a decision which was informed and based on what was important to them and this was achieved by ensuring that the risk communication tool was balanced, accurate and up to date.

The purpose of value clarification in the process of shared decision making is to help patients recognise that their values play an important role in decision making. It also helps patients to determine which treatment option and its attributes are important to them and share their values with the healthcare professionals and their families [28].

In the PANDAs study, an explicit "balanced technique" was used to help patients compare and indicate the relative importance of the treatment option as well as the attributes of each option [17]. However, patients found it difficult to answer the "negatively phrased" questions (e.g. `how important is it to avoid weight gain?’), especially when the response was also negative (eg `not important’). This difficulty was compounded when they were required to rate the importance (from `not important’ to `very important’). As a result, the questions asked in the value clarification section of the PtDA were modified and the response items were reduced to a yes/no option. It was clear that the accuracy of response to the questions in the PANDAs PtDA needed to be weighed against the ease of answering the questions.

The PtDA was delivered in a booklet based on feedback from the patients, healthcare professionals and expert panel during the needs assessment study. PtDAs can be developed in other formats including web-based compact discs or digital versatile disc; audio book, common decision board and patient focus groups. Increasingly web-based decision aids are preferred by developers. A paper-based PtDA was chosen because both healthcare professionals and patients during the feasibility study found the paper-based PtDA acceptable and the PtDA itself needed to be used by healthcare professionals and patients together in a consultation. In addition, a paper PtDA was portable and could be taken home as required by the patient for further consideration and it does not require any equipment. Most of the target population for the use of the PtDA were above 50 years of age and the assumption which was made for the study was that the majority of people in this age group prefer to use paper-based tools and are generally less familiar with using computers or the internet. However, this preference is changing [29]. A survey conducted among orthopaedic surgeons in the UK found that more than half of the respondents preferred a booklet as a PtDA to other media such as web-based PtDA, CD/DVD and audiotape [30].

However, the disadvantages of using a paper-based PtDA are that the latter requires printing and mailing which may hinder wide dissemination and printing and reprinting of the PtDA was costly. By comparison, developing a web-based PtDA often needs only a single payment for computer programming and web design in the initial phase and maintenance and updating of a web-based PtDA requires less cost than reprinting of the paper-based PtDA.

Finally, web-based or computer-assisted PtDAs allow data to be collected and, if needed, summarized - so that it can be fed back to the patients and healthcare professionals, an advantage which paper-based PtDAs do not have. Nevertheless, a patient decision aid, regardless of the mode of delivery, only serves as a tool to facilitate discussion on the decision; it is not meant to replace a clinical consultation.

This study has demonstrated the feasibility of the IPDAS and the UKMRC frameworks for the development and evaluation of a PtDA. The IPDAS guideline was useful for informing the content, development and formatting of the PtDA, while the UKMRC framework helped to link the process of development with the evaluation and implementation of a complex intervention. Both frameworks are complementary and essential components to guide the development and evaluation of future PtDAs.

CJN conceptualised the study, designed the methodology, conducted the study, analysed the data and drafted the manuscript. NM conceptualised the study, designed the methodology, analysed the data and helped to draft the manuscript. AB conducted the study, analysed the data and helped to draft the manuscript. BC conducted the study, analysed the data and helped to draft the manuscript. All authors read and approved the final manuscript.