Impact on patients of expanded, general practice based, student teaching: observational and qualitative study

Quantitative

We used the patient enablement index (PEI)6 and the consultation satisfaction questionnaire (CSQ)7 to assess patients' enablement and satisfaction (see bmj.com). The patient enablement index contains six questions, each with a score range from 0 to 2. The consultation satisfaction questionnaire contains 18 mixed positive and negative statements, scored on a 5 point Likert scale ranging from “strongly agree” to “strongly disagree” and grouped into four different components. Higher scores indicate higher levels of enablement and satisfaction. We combined the scores of the components of the consultation satisfaction questionnaire to indicate overall satisfaction as described elsewhere.6 8 We expressed scores for both instruments as a percentage of the maximum score obtainable (12 and 90, respectively).6 8 9

Between March and July 2003, we gave both instruments to all patients attending teaching consultations and to all patients attending selected non-teaching consultations with the same general practitioner at a similar time of day. We excluded visually impaired, frail, or severely ill patients and those under 16 years.

On the basis of previous studies, we calculated that for each general practitioner teacher, a minimum of 28 teaching consultations and 28 non-teaching consultations would permit detection of a mean difference of 25% in the score of the patient enablement index (assumed standard deviation 32.5%), and 3% for each of the consultation satisfaction questionnaire components (assumed standard deviations 3.8%, 3.5%, 4.3%, and 4.0%; 80% power, 5% level of significance).6 7

We aimed to include at least one general practitioner teacher from each practice whose number of consultations fulfilled these power requirements. Several general practitioner teachers' consultation numbers fell just below this threshold, but six teachers (including at least one from each practice) saw sufficient patients to detect differences as above, with a power of at least 65% (table 1).

Age and sex of patient, and practice have previously been found to influence the scores of both instruments.7 9 We examined their influence at the level of all patients and at the level of individual general practitioners.

At the level of all patients, to account for the nested structure of the data, we fitted multilevel linear mixed effects models allowing for random effects of practice and general practitioner within practice. This used the method of restricted maximum likelihood estimation to compare scores in the teaching and non-teaching groups, adjusting for age of patient.7 10^–12 We omitted sex of patient from the final model as we found it not to influence either score significantly.

At the level of general practitioners, we used Student's t test to compare the scores of both instruments in teaching and non-teaching consultations.6 We also used the non-parametric Mann-Whitney U test to analyse the scores of the patient enablement index because of concern about whether these scores were normally distributed. The results were similar to those from the t test. We then used multiple linear regression models to compare scores of both instruments in teaching and non-teaching consultations, adjusting for age of patient. We found no major violations of modelling assumptions in the linear regression or multilevel models.

We considered P values of less than 0.05 to be significant. We used SPSS, version 12.0 (SPSS Incorporated, Chicago, IL, USA) and R, version 2.0 (www.itc.virginia.edu/research/R/) for our analysis.13

Qualitative

In each practice, during the period of expanded teaching, we identified patients attending prearranged teaching sessions, together with matched patients not participating in these. We invited a selected number of each type of patient to attend separate focus groups. We selected patients to achieve maximum diversity with respect to age and sex. Between June and December 2003, 10 focus groups took place, two from each practice, with 34 patients who had attended prearranged teaching sessions and 20 patients who had not.

We gave all discussion groups three topics drawn from the literature: potential impact on service delivery, experience of students, and attitudes towards increased community based teaching of students.2 3 14 We used a nominal group technique in the initial teaching and non-teaching groups to allow participants to rank issues in order of importance. We presented specific issues emerging from the nominal technique to subsequent focus groups. We adopted a flexible approach to allow emerging areas to be explored and used specific prompts to raise issues identified in other studies if they did not emerge during open discussion.

We recorded discussions on tape, transcribed them, and used Atlas.ti software, version 4.1 (Scientific Software Development, Berlin, Germany) to analyse them. With the study's objectives in mind, four of the authors (JB, TQ, JE, and AH), independently examined two early transcripts to identify emergent themes. As the project proceeded, focus group discussion was informed by data analysis. One transcript was then coded independently by the same authors and an initial code list derived. This code list was subsequently tested against a later transcript, leading to its adjustment. The authors then independently applied the adjusted code list to two further transcripts. We discussed differences in code application, and, where needed, we revised the code list. After this, TQ applied the final code list to all transcripts.

The selection of relevant qualitative findings was informed by a combination of frequency of comment and level of consensus on importance within each focus group. We sent a summary of the findings to the 54 participants. Thirty nine responded, of whom 38 felt that their views had been clearly represented and one felt partially represented.

Analysis by all patients

The mean score on the patient enablement index was 33.4 (SD 30.4) for patients attending teaching consultations and 31.2 (SD 28.9) for patients attending non-teaching consultations. The corresponding statistics for score on the consultation satisfaction questionnaire were 80.4 (SD 10.1) for “teaching patients” and 78.5 (SD 10.5) for “non-teaching patients.”

For all patients, we found no difference in respect of patient enablement index scores (table 3). Comparison of mean scores for the consultation satisfaction questionnaire with the linear mixed effects model provided weak evidence that patients attending teaching consultations were more satisfied than patients attending non-teaching consultations, but this difference was just below statistical significance (table 3).

Analysis by general practitioner teacher

Analysis at the level of general practitioner teacher, adjusting for age of patient, showed no significant difference between teaching and non-teaching patients for either instrument (table 4).

Impact on service delivery

We had no specific reports from participants of being less able to see their preferred general practitioner as a direct result of teaching activities. Many patients preferred to see their “own” doctor. Most accepted that it was not always possible to do this, but few reported being given reasons why their general practitioner was unavailable when requesting an appointment. Patients said they would prefer such information and felt teaching was an acceptable explanation for absence (quotation 1, box). However, continuity of care from the same general practitioner was particularly important for patients with chronic conditions or where trust was crucial (quotation 2, box).

Patients' willingness to participate in teaching

Patients reported more advantages than disadvantages of participating in prearranged teaching sessions and teaching consultations. Advantages included improved consultations resulting from more time and perceived greater depth of consultation, a better understanding of their condition and treatment, and feelings of improved self esteem and altruism where patients felt that they had done something worthwhile or given something back to the NHS. Some patients also had a perception that their general practitioner was more up to date as a result of participating in teaching. The only direct disadvantage seen was the increased time allowed for teaching consultations and its possible knock-on effects for other patients.

Teaching and non-teaching patients alike said that they were willing to be seen by students. However, this willingness was conditional and based on several aspects such as the nature of their complaint or possible examination and the sex and number of students likely to be present. Such concerns were not confined to women (quotation 3, box).

Worries were also expressed about students' competence. However, unlike in other studies, patients did not express concerns about confidentiality.

Informed consent

Two interrelated aspects of consent emerged: the specific information given to patients before seeking their consent and the point at which consent is sought. Participants wanted to know the sex and number of students, although this information was rarely provided. Information about students' competence and experience and the likely nature of the consultation or examination were also felt necessary for truly informed consent.

The timing of consent was felt to be equally important. Participants attending teaching consultations expected to be informed well in advance—for example, when they first telephoned to make an appointment. In a few isolated instances patients were informed only when they reported to reception and others only when met by the general practitioner on their way to the consulting room. Although these instances had not created problems for the patients, situations where consent was sought late were thought inappropriate, leading to potential embarrassment associated with the reason for the visit or feeling pressurised to consent at short notice (quotation 4, box).

Primary care as the patients' territory

Many patients distinguished between what “has to be accepted” in hospitals, in terms of student presence, and what is acceptable in the general practitioner's surgery (quotation 5, box). Perceived lack of personal power and space (quotation 6, box), and the more urgent need for treatment in a hospital context (quotation 7, box), were contrasted with the more intimate, ongoing relationship with the general practitioner and personal setting of the surgery (quotation 8, box).

Comparison with other studies

Quantitative analysis showed that mean scores and standard deviations recorded for the patient enablement index and consultation satisfaction questionnaire resembled those seen in other studies of a similar size that used the same approach to both measures.6 Overall patients' satisfaction was therefore similar to that observed previously elsewhere.6 8 9

The results of the qualitative analysis support findings of other studies indicating the benefits that patients derive from involvement in teaching: longer and more thorough consultations,16 17 better understanding of their complaint, and improved self esteem.2 3 In contrast to previous studies, patients did not express concerns about confidentiality, exploitation, or reinforcement of a stereotypical “sick” role.3 14 Although patients value continuity and, as noted elsewhere,18 this may be crucial for some, patients in this study reported no direct problems of continuity.

Strengths of the study

This study was new in combining quantitative and qualitative methods and in exploring the views not only of patients participating in teaching but also of those who were not. Participants in the survey and focus groups were older than registered patients, and survey participants were more often female. This bias may reflect a similar bias in consultation patterns, meaning that findings may be generalised, at least to those attending surgeries.

Limitations of the study

We attempted to measure the impact on availability and access. However, changes to booking procedures and new, nationally driven triage systems meant that this could not be done on a consistent basis.

Informed consent in a general practice setting

As previously reported, patients expressed concerns about the sex, number, and competence of students and about the likely nature of any examination.3 14 17^–19 Information about these are key components to enable patients to give informed consent to teaching. In this respect, the study expands and clarifies issues surrounding informed consent.20 21 For patients, these issues are framed within a perceived difference in the way that they see the practice environment, compared with hospital. They speak of the practice as being more legitimately within their own control than is the case in hospital.

Conclusions

This difference in perception has important implications for the range and volume of teaching that may be acceptable in primary care without undermining the relationship between general practitioner and patient. The extent to which general practitioner teachers are sensitive to these subtle differences and their effect on patients' expressed wishes for information before giving or withholding consent for teaching may be crucial to maintaining doctor-patient relationships in teaching practices. When seeking to involve students in consultations with general practitioners, respect for these wishes and recognition of the need to reaffirm these actively with patients is likely to be important for the successful expansion of community based teaching in the long term.

Amendment

This is Version 2 of the paper. In this version, the word “quantitative” in the title has been corrected to “qualitative.”