Abstract
Background: Although a ‘patient-centred’ approach to general practice consultation is widely advocated, there is mixed evidence of its benefits.
Aim: To measure the costs and benefits of using a prompt to elicit patients' concerns when they consult for minor illness.
Design of study: An open randomised controlled trial.
Setting: Four training semi-rural general practices in the south-east of the United Kingdom.
Method: Patients identified during the first part of the consultation as having a self-limiting illness were randomised to a second part of the consultation that was conducted ‘as usual’ or involved a written prompt to elicit the patient's concerns. After each consultation the doctor noted the diagnosis and the consultation length and the patient self-completed a questionnaire containing measures of satisfaction, enablement and anxiety.
Results: One hundred and ten patients were studied. Patients in the elicitation group reported a small but significant increase in the ‘professional care’ score of the consultation satisfaction questionnaire (88.2 versus 80.9, mean difference = 7.3, 95% confidence interval = 2.0 to 12.6) but no other benefits were detected. Consultations in the elicitation group, however, were longer by about a minute.
Conclusion: Given the pressures on consultation time in general practice there must be questions about the practical value of eliciting patients' concerns if the benefit of doing so is small and the cost large.
Introduction
TWENTY years ago, when Tuckett and his colleagues studied over 400 general practitioner (GP) consultations they found that in only 6% did the doctor make an active effort to elicit the patient's view about ‘diagnostic significance’, in 3% about ‘treatment action’ and in none about ‘implications’.1 More recently, in an investigation of the MRCGP (Member of the Royal College of General Practitioners) video examination involving study of ‘best practice’ videos, it was again found that the three criteria demonstrating patient-centredness were rarely achieved.2
Elicitation of the patient's view enables the GP to understand the illness from the patient's perspective, as opposed to the biomedical construct of disease. The benefits of such a strategy include relieving patients' anxieties3-5 as well as improving satisfaction and enablement.6 However, the costs of this form of clinical practice, particularly the time involved, are, for the most part, forgotten in the rush to identify and quantify benefits. For most GPs, time in the consultation is the most limited resource and a more patient-centred approach might well cost more than the benefits it confers. A study was therefore carried out that involved randomising patients with acute self-limiting conditions to a policy of ‘usual practice’ or to a deliberate attempt to elicit concerns to assess the overall value of the approach.
Method
GP registrars attached to three vocational training schemes in the south-east of the United Kingdom and their trainers, were invited to participate in the study. GP registrars were included so that, if the chosen intervention was found to be effective, it could be used in training. Ethical approval was obtained from the three relevant local research ethics committees: Brighton, Haywards Heath and Worthing.
Patients presenting with self-limiting conditions were randomised to the intervention and control groups. The decision whether or not to include a particular patient was taken after the history and examination, when the doctor was able to judge that the problem was a self-limiting illness. This incl-usion strategy involved asking for consent from a much larger group of patients, sometimes all the patients attending a surgery. Patients were informed that the aim of the study was to examine how different ways of providing explanations affected patient satisfaction. Patients were excluded if they were not judged to have a self-limiting illness, if they were to be referred to hospital or given a prescription other than for symptom control, or if they spontaneously expressed a clear concern about their illness.
If the patient was eligible for inclusion the top sheet from a randomly arranged pile was turned over. If the underside said ‘control’, the consultation proceeded as normal; if ‘intervention’, it included written prompts (Box 1) to facilitate elicitation of concerns.
HOW THIS FITS IN
What do we know?
Eliciting the patient's own view of their illness is widely advocated as a characteristic of a good consultation, although evidence for the benefits is mixed.
What does this paper add?
The benefits of eliciting the concerns of patients with self-limiting illness are relatively small but the costs, in terms of time, are significant.
Box 1. The written prompt.
‘May I ask if you have any concerns about this “ … ” (illness/pain) you have come about today?’
followed by:
‘Anything in particular about the “ … ”?’
and, if still unforthcoming:
‘What is it about the “ … ” that concerns you?’
In order to validate the intervention, 14 consultations, five intervention: two from one trainer and one each from three different GP registrars, and nine controls: three from one trainer, five from one GP registrar and one from another GP registrar, were recorded by audiotape or videotape. These were presented to an independent blinded assessor to judge whether the doctor had actively elicited the patient's concerns: 13 were assessed correctly and one control was marked ‘don't know’.
At the end of the consultation the doctor recorded the consultation length and the diagnosis made. Participating doctors were instructed to estimate the consultation length to the nearest minute using a clock to note the time at the start and end of the consultation. The patient (or accomp-anying adult if the patient was under 16 years old) then completed a questionnaire in the waiting room that contained the consultation satisfaction questionnaire (CSQ),7 the patient enablement instrument,8 and the six-item short-form of the state scale of the Spielberger State-Trait anxiety inventory.9 The professional care component of the CSQ was chosen as the main outcome measure, as the questions most closely matched the desired consequence of eliciting patients' concerns; the questions that comprise this component of the CSQ are given in Box 2. The scores of the three questionnaires were adjusted to fit scales of 0–100 for the CSQ and patient enablement instrument and 20–80 for the anxiety measure.
Box 2. Professional care component of the consultation satisfaction questionnaire.
How much do you agree with the following statements?
2. This doctor was very careful to check everything when examining me.
3. I will follow this doctor's advice because I think he/she is absolutely right.
6. This doctor told me everything about my treatment.
9. This doctor examined me very thoroughly.
10. I thought this doctor took notice of me as a person.
12. I understood my illness much better after seeing this doctor.
13. This doctor was interested in me as a person.
A sample size was calculated using the standard deviation of 3.5 for the professional care component of the CSQ from previously published data.10 It was calculated that a sample size of 120 would be needed to show a difference of 1.7 points on the scale at a power of 80% and significance of 0.05. Data from the questionnaires were analysed using SPSS. Missing values in incomplete questionnaires were replaced by the series mean of the relevant variable.
Results
Six GP registrars and two trainers from four practices partic-ipated and recruited 81 patients and 29 patients respectively, 54 in the control group and 56 in the intervention group. The reasons why patients were excluded from the study were noted for 129 consecutive patients seen at one practice. Eighty-nine (69%) did not have a self-limiting illness, 20 (16%) had not receive a consent form, 4 (3%) refused consent and one had already participated in the study. None of the 129 patients spontaneously expressed a concern. Figure 1 shows a flow diagram of the recruitment process using this sub-group data to model the likely numbers from which the final sample was obtained. The range of diagnoses in the patients included in the study is given in Table 1.
Figure 1 Modelled recruitment flow chart.
Table 1 Diagnosis in study patients.
The score for the professional care component of patient satisfaction was 80.9 (standard deviation [SD] = 16.1) for controls and 88.2 (SD = 11.8) for intervention patients, mean difference 7.3 (95% confidence interval [CI] = 2.0 to 12.6). Sub-group analysis of patients seen by just the registrars found similar results, with a professional care score of 82.0 (SD = 14.8) for controls and 89.0 (SD = 11.6) for intervention patients, giving a mean difference of 7.0 (95% CI = 1.2 to 12.8). There were no significant differences in any of the other outcome measures (Table 2).
Table 2 Results of all outcome measures.
Intervention consultations were on average 1 minute longer than controls, that is 11.0 minutes versus 10.0 minutes, although this difference did not reach statistical signif-icance. The length of the consultation (in minutes) and the intervention status of the patient (intervention or control group) were then entered into a multiple regression to assess their ability to predict patient satisfaction with professional care. The results showed that, while length of consultation (β coefficient = 0.21, P<0.05) contributed less than the intervention status of the consultation (β coefficient = 0.29, P<0.005), it was still a major predictive factor.
Discussion
Summary of main findings
This study has shown that patients with acute self-limiting illness are more satisfied (on one of the measures) when GPs are prompted to ask them about their concerns. Yet this benefit was only achieved at the cost of a 10% increase in consultation time (which itself seemed to be responsible for some of the discovered benefit). Moreover, the benefit does seem rather meagre. No impact was discernible on patients' reports of general satisfaction, quality of the relationship with the GP, enablement or anxiety. Perhaps a larger study might have uncovered some changes in these measures, but their effect would have been small. On the other hand, a 10% increase in consulting time adds up to a lot of GP resource at a time when the latter is under severe strain.
Strengths and limitations of this study
The study was designed to identify a difference of over 1.7 points on the professional care scale of the CSQ. However, the study found that the SD of the scale's scores was considerably higher than that used in the power calculation (the published data turned out to have been erroneous10) meaning that the study only had power to detect differences of about 7 points or more. Smaller differences in the other scales may have been significant with a larger sample size, although the main conclusion of the study, that investment of consultation time achieves relatively small benefits, still stands.
Only patients presenting with self-limiting illnesses were studied. The management of self-limiting illness was chosen for this study partly in response to qualitative work demonstrating dissatisfaction with doctor–patient communication in such consultations,3,11 and because it is in this context that patients' and doctors' views of the illness are most likely to diverge. The doctor may well believe ‘there is nothing to worry about’, while the patient remains concerned. Much of general practice, however, involves chronic illness and these findings cannot be generalised into that area. The results are based on a small number of doctors and patients. It is poss-ible that in other populations more significant benefits can be achieved, but the increased cost in terms of consulting time seems intuitively right: eliciting — and then listening to — concerns is an additional task in the consultation and must therefore involve expenditure of more time.
Comparisons with existing literature
Patient-centredness remains an aspiration for good general practice and determining the patient's perspective is a part of the accepted strategy.12,13 There are, however, challenges to this perspective, arguing that this approach is not what patients want, nor expect, from their GP.14-16 Yet this debate presumes that elicitation of views is cost-neutral, that the merits of the strategy can be determined by examining patients' wishes and clinical outcomes. The important question, surely, is whether the benefits outweigh the costs, in this case the additional time required.
Implications for future research or clinical practice
The extra time devoted to eliciting patients' views could be justified if time was ‘saved’ by reductions in patients' future consulting behaviour. This study did not examine this outcome — although it may merit empirical investigation. It is also possible that the enthusiasm for eliciting patients' views has been determined by the satisfaction it gives GPs themselves rather than for any major benefit for patients. However, in the absence of evidence for these alternative advantages, judgement of the value of eliciting concerns must involve a trade-off between time spent and benefits gained within the consultation. According to a recent study2 it is not a skill demonstrated by new recruits to general practice and perhaps it is not something that should be further emphasised if it is as costly as this study suggests. The ideal may be desirable, but at a time when the fundamental problem of British general practice is a shortage of consultation resources, the best may be the enemy of the good.
Acknowledgments
This study derives from an MSc in General Practice carried out at Guy's King's and St Thomas' School of Medicine. We would like to thank the GPs and patients who took part in this research.
- Received March 31, 2003.
- Revision received June 6, 2003.
- Accepted December 23, 2003.
- © British Journal of General Practice, 2004.