Patient "self management" or "self care," a concept that enables patients to take a guided but ultimately personal involvement in the management of their condition, is an increasingly debated element of healthcare provision. It is particularly relevant as the prevalence of long term conditions increases and growing numbers of people desire a more active role in their own care with a less paternalistic approach from healthcare professionals 1. Effective self care has the potential to improve clinical outcomes and reduce use of healthcare resources 12.

Asthma is an ideal condition in which to strive for improved patient outcomes by optimising self management because it typically fluctuates over time, with symptoms and exacerbations that can potentially be minimised with self monitoring and appropriate adjustment of treatment 34. Self management of asthma is currently suboptimal in many patients, with around 50% self managing in ways that differ from recommended guidance 567.

A key step in improving the self management of asthma is to understand what patients consider important. Patient education programmes designed to improve self care have traditionally centred on what health professionals consider to be important, for example, lung function, asthma symptoms and bronchodilator use in asthma 8. Previous research has shown that patients have different perceptions of asthma compared to health professionals and that education tailored to meet patients' perceptions is more likely to change behaviour 7.

This study was designed to quantify the relative importance of features of the management of asthma from the patients' perspective. We used discrete choice experiment methodology, a type of conjoint analysis that has been shown to be a rigorous survey technique for eliciting preferences 9. It is increasingly being used to identify patient and public preferences for health care 1011. The technique allows respondents to choose their preferred option between hypothetical scenarios designed to reflect the different attributes that real world decisions would contain, and to make trade offs between these attributes to reveal their preferences. This technique of revealing preference through choice is a truer representation of real life decision-making and as such may be a better tool for establishing preference than data based on the ranking or rating of individual components of asthma management 12.

A clearer understanding of such preferences may help healthcare professionals tailor an acceptable personalised management of asthma with their patient and consequently move nearer to controlled asthma 13.

A brief description of discrete choice experiment, a working example and a glossary of technical terms and jargon are provided at Additional file 1.

We carried out a discrete choice experiment (DCE) to determine the characteristics of long term asthma management that patients consider most important, requiring them to make choices between hypothetical scenarios and thus reveal their preferences.

Ethical approval was granted by Warwickshire Local Research Ethics Committee on behalf of COREC UK reference number 04/Q2803/66.

A total of 148 questionnaires were returned after one reminder, of which one was returned blank, from a total of 348 sent out, giving a useable response rate of 43%. Table 2 summarises the sociodemographic characteristics of the study population, while Table 3 compares basic characteristics of responders with non-responders.

Non-responders, by definition, did not consent to their involvement in the study. Consequently, a more detailed comparison of characteristics or features of responders and non-responders was not possible.

The Relative Importance results are presented in Table 4.

The outputs shown throughout are the means of the parameters calculated at the level of individual respondents. The means at aggregate level demonstrating relative importance were also calculated and are included for comparison, displaying where a non-homogenous response occurs. See Additional file 1.

Figure 2 shows the importance that respondents placed on changes between different levels within-attributes. The degree of importance is seen by changes in utility values for levels within the attributes. All 11 successive within attribute transitions were statistically significant (P < 0.05), with the single exception of changing the management of an asthma crisis from 'yourself' to 'visiting a GP/nurse'. Those changes with the highest relative negative impact on respondents' views of their asthma management were:

• changing from 'no more than 2' to '3' inhalers,

• change in steroid dose from 'low but high when needed' to 'always high',

• being encouraged 'to visit a hospital for crisis management' rather than being encouraged 'to manage yourself' or 'attend the local GP surgery',

• symptom relief provided by current treatment changing from 'completely' to 'mostly',

• changing from 1 to 2 inhalers.

The study emphasises the importance of keeping treatment regimens simple. The results showed that adults with moderate or more severe asthma considered that a simple treatment regimen was the most important consideration in the long-term management of their condition, rather than symptom control without compromise. For example, two of the top five highest utility shifts between levels related to the number of inhalers they needed to use. Changing from 'no more than two' to 'three' inhalers had the highest relative negative impact on respondents' views of their asthma management. While noting the caveats of the relative importance analysis, number of inhalers was ranked the most important attribute of asthma management at both the aggregate (29.3%) and individual levels (21.9%), suggesting a reasonably homogenous view.

This preference for simpler treatment and fewer inhalers confirms in a more systematic and rigorous way preferences for "fewer drug treatments" and "just one inhaler" reported in a previous pan-European study 7 and confirms the findings from patient interviews in our previous study 16. Asthma is only one part of people's lives and treatments that may need to be taken for decades should be offered in the simplest format. Willingness to pay from the patients' perspective – another factor that may influence treatment preference – was not addressed in this study; the cost of therapy to patients may be less important in the UK than in other healthcare settings; it was not rated highly as an issue in our qualitative study 16.

The factor that patients rated as being of next highest relative importance, and which had the second greatest utility shift, was the dose of inhaled steroid. Scope for lowering the steroid dose without loss of asthma control has previously been described 18 and the addition of an inhaled long-acting beta agonist often permits better control and use of a lower dose of inhaled steroid 14.

Use of a personalised asthma action plan came next in patients' ranking of relative importance of the attributes of asthma management that they were asked about. A discouragingly small number, only 12 (8%) of respondents, indicated that they held a written personalised asthma action plan – two centres each accounted for three of these patients and a further six practices each had one patient with a plan. This low number of patients with an action plan is similar to that found in previous studies 19 and is disappointing, especially because it has previously been shown that even those without plans would feel comfortable adjusting therapy themselves 16. Written asthma action plans have been shown both to improve outcomes 3 and to improve compliance with asthma therapy 20, to be cost-effective 21 and are strongly recommended in asthma guidelines 14. It may be that lack of familiarity with the nature and benefits of using a personalised plan, by both medical professionals and patients, may have influenced these results and that a greater knowledge would increase the popularity and use of what may be the single most important non-therapeutic intervention in asthma management. In this study, patients indicate a desire for "brief" rather than "full" written instructions.

The next ranked factor was asthma crisis management. The utility analysis showed that patients preferred to avoid attending hospital even in the event of a crisis, a theme we have reported in a different disease area and population 22. Knowledge of patient preference can inform the clinician but will not, of course, be the only factor to consider when deciding how and where to manage an acute exacerbation of asthma.

Perhaps surprisingly, controlling asthma symptoms was ranked lower in patients' ranking of importance, and relief of symptoms was considered least important in the range of attributes tested. However, this does not mean that people with asthma do not consider symptom relief important, but indicates that respondents considered it less important than the other attributes of asthma management they were asked to rank. This suggests that patients were prepared, at least to some extent, to trade off elements of efficacy for what they perceived to be other benefits, such as lower doses of inhaled steroids.

Both 'asthma crisis management' and 'controlling your asthma symptoms' had higher relative importance statistics when determined by the individual level method than by the aggregate level method. This means that there was a division of opinion within respondents as to which level in each of these two attributes was the most desirable.

There was some variation between respondents in the extent to which they wanted to manage their own asthma symptoms. Nearly two-thirds put a higher utility value on being encouraged to 'change your own therapy' than 'speak to a doctor or nurse before making changes to treatment' in the attribute of controlling asthma symptoms. This indicates a split between patients wanting a collaborative/active role in making changes to their asthma therapy and those wanting a more passive role, at a similar level to that reported previously 23.

One of the greatest strengths of this study is the use of discrete choice experiment methodology, which is a rigorous method of eliciting preferences. Previous studies have demonstrated that respondents tend to behave in an internally valid and consistent manner when answering DCE questions 24. The study explored patients' preferences between only the attributes and levels that were offered, but these had been identified as being important from patients interviewed in a previous study 16. The majority of the respondents found the questionnaire easy to complete, although it is possible that the type of questionnaire and the task, which is likely to have been unfamiliar to recipients, influenced the overall response rate.

Another possible limitation to the study is that the majority of respondents were female (65%) and aged over 55 years (48%). However, this is similar to previous studies exploring adult asthma patients' attitudes to their treatment 923. Responders were generally older (mean age 55 years) than non-responders (mean age 45 years) (P < 0.01), but there was no statistically significant difference in gender between respondents and non-respondents (P = 0.3%).

Taking a flexible, patient-centred approach to asthma management means focusing on issues that patients consider important. Our study indicates that this means making treatment as simple as possible, with as few medications and inhalers as can achieve symptom control – ideally fewer than three, or even two, inhalers. It also means using the lowest dose of inhaled steroid that can effectively control asthma and avoiding hospitals for emergency care, as well as minimising asthma symptoms. There is clearly room for improvement in increasing the number of patients receiving personalised asthma management plans, which should improve outcomes by increasing compliance.

John Haughney has received fees from AstraZeneca, Boehringer Ingelheim and Merck Sharp & Dohme for speaking at meetings and from AstraZeneca, GlaxoSmithKline, Merck Sharp & Dohme, Novartis and Schering Plough for consulting. Monica Fletcher has received reimbursement from AstraZeneca, Schering Plough, Altana and Novartis for attending international conferences, fees for speaking at meetings and funds for research. Stephanie Wolfe has received fees from Merck Sharp & Dohme for staff training, Novartis for consulting and AstraZeneca, GlaxoSmithKline and Merck Sharp & Dohme for speaking at meetings. Julie Ratcliffe has no competing interests. Roger Brice has no competing interests. Martyn Partridge has received fees for lecturing from AstraZeneca, GlaxoSmithKline, Novartis and Vitaris. He has in addition received sponsorship to attend European Respiratory Society and American Thoracic Society meetings from AstraZeneca and Novartis, and the Astra Foundation have supported a non-clinical Research Fellow.

JH, MP and MF conceived and designed the study, interpreted the data and wrote and reviewed the manuscript. SW supervised the processes of the study, interpreted the data and reviewed the manuscript. JR helped conceive and design the study, interpreted the data and reviewed the manuscript. RB carried out the analyses, interpreted the data and reviewed the manuscript. All authors read and approved the final manuscript.