This study is part of a larger project funded by the Andalusian Council and analyses the communicational profile of medical tutors and senior resident physicians of family and community medicine in four Andalusian cities.17

This is a cross-sectional, multicentre study carried out in health centres in four Spanish provinces. Andalusian physicians of family medicine participated, accredited as tutors and residents in their last year of training.

They were videotaped during regular working hours. The main researcher then selected 4 video recordings of each professional (excluding their first consultation and bureaucratic or internal consultations) at random, which were subsequently analysed by two people alien to the research and trained in the use and management of video recordings using the CICAA survey.

The intraobserver agreement by the intraclass correlation coefficient was 0.89 (95% CI 0.79–0.94) for the first observer and 0.91 (95% CI 0.84–0.95) for the second. The interobserver agreement of both was 0.94 (95% CI 0.89–0.96).

The main dependent variable was the total consultation time, measured in minutes, from the moment the patient entered the consultation until he left.

It was divided into three parts:

• a)

  exploratory phase, from the beginning of the interview to the physical examination.

• b)

  physical examination length.

• c)

  resolution phase, from the end of the physical examination until the end of the interview.

The main independent variable was the score obtained on the CICAA17 questionnaire with three possible values for each item: 0 points (did not accomplish the set tasks), 1 point (acceptably accomplished the set tasks) and 2 points (achieved the set tasks completely or almost completely). The score can range from 0 to 58 points. In addition, two new variables were established: patient-centred communication and patient participation in decision making.

• -

  Patient-centred communication: If a professional obtained a score of 7 points or more in the sum of the items 5, 10, 11, 12, 13, 14, 15, 16, 18 and 26 of the questionnaire, the communication was considered patient-centred (corresponding to the mean and median of the variable in this sample). Numbers relating to exploring the psychosocial sphere and patient-centred consultations in validating the questionnaire.19

• -

  Patient involvement in decision making: The patient was considered to have been involved in decision making when the score in number 25 was 1 or 2, according to the work published by Moral et al.7

Other independent variables were the personal and professional characteristics of physicians that emerged during the interviews.17

The research project was approved by the Ethics Committee of Health Research in Referral Hospitals. The informed consent of healthcare professionals and patients was obtained in writing or verbally during the video recordings.

Statistical analysis was performed using the program SPSS (15.0), considering a level of less than or equal to 0.05 as significant.

The analysis of consultation length according to the characteristics of the physician, the patient and the interview, was conducted by using Student t tests, ANOVA, correlation and/or simple linear regression. Finally, a multiple linear regression model was constructed with length as the dependent variable, having previously checked the conditions.

There were 476 video recordings in total, excluding 40 due to technical difficulties. Of the 119 participants, 60 were residents (50.4%) with 229 interviews (52.5%) and 59 tutors with 207 interviews. The characteristics of physicians and interviews can be seen in Tables 1 and 2.

The average duration of consultations was 8.8min (sd: 3.86), with a median of 8min. Almost double the time was taken for the resolution phase than in the operative phase (Table 2).

There are slight differences in consultation length between tutors and residents; however, tutors took longer in the exploratory phase of the interview, and residents took longer in the operative phase (Fig. 1).

Figure 1.

Consultation time distribution according to the type of professional.

(0,1MB).

In the bivariate analysis, a longer consultation length took place in urban centres, with younger professionals, those who had a patient-centred profile, when there was shared decision-making, in consultation reviews and when the patient was accompanied (Table 3).

A correlation between duration of the consultation and the number of reasons for the consultation was also found (r=0.353; p<0.001), with the professional's score obtained from the CICAA scale (r=0.420; p<0.01).

By contrast, consultation time was not related neither to the type of problem, neither with the gender of the doctor, the patient, nor with the doctor–patient gender interaction.

In the multivariate model, longer consultations were associated with better CICAA scores, a greater number of reasons for the consultation, if the patient was accompanied, the health centre type (urban), and less daily consultations (Table 4). The coefficient of determination (r2) was 0.32.

Before analysing the results, we must discuss possible limitations including voluntary sample selection. Although this is the method most frequently used in video-recorded studies20 is not exempt from possible selection bias, and may represent a subpopulation of more motivated professionals with a better communication profile.

Another difficulty would be behaviour change whilst being recorded; however it has been shown that this did not occur.20 In our work, recording a minimum of 1h is recommended, selecting multiple consultations of each professional and eliminating the first.

Regarding scale of measurement, there are many questionnaires to measure patient-physician communication, but only two that have been validated in Spain, one is the CICAA scale, which has certainly demonstrated its validity for this purpose.18,19

It has also been possible to produce a residual confounding bias as not having measured other variables probably related, among them, the reason for consultation.

Finally, this type of design of research studies does not allow establishing causal relationships between variables.

The average consultation length in this study is around 9min, similar to that obtained by other authors in Spain.10,21 At European level, the average length is between 10 and 15min; according to Deveugele,4 shorter consultations occur in Germany and Spain with 7.6 and 7.8min respectively; in England and Holland duration is around 10min and in Belgium and Switzerland around 15min.

For some authors, consultation length should be around 20min to achieve an approach that includes encouraging a healthy lifestyle, detecting risk factors (hypertension, alcoholism, smoking), improving patient and professional satisfaction.1,2

There is no difference in the total consultation time between tutors and residents, although they distribute time differently, tutors spend more time on the exploratory phase, and residents are quicker in the anamnesis.

The bivariate analysis shows how older doctors perform shorter consultations, which was observed in resident physician training.22 However, after performing a multiple regression analysis, like other authors we have not found a relationship between consultation length and age of the physician.4

No differences in consultation length in relation to the gender of the professional were detected, unlike in other areas, where female medical staff took 2min longer than men, involving the patient in the consultation more.5,12–14 Although there are publications that describe an increase in quality of care and communication skills without increasing the consultation time,23,24 in this study, as in other studies,6,7,25,26 better communication (a higher CICAA score), patient involvement in decision-making or developing a patient-centred approach are factors that lengthen the consultation. It is inevitable, as several authors claim, that listening to the patient, and involving them in the process, would require more time.6,7,25,26

In urban centres, consultation length is increased, possibly because patients had more reasons for the consultation.4 It has been found that the more reasons the patient has for the consultation, the longer the duration due to the type of diagnosis the physician must give.4,5,8,27 The consultations due to psychological,4 gynaecological,5 social and psychosocial problems are usually longer than strictly biomedical consultations.8,28

As Hutton and Gunn8 noted, follow up consultations related to the elderly or patients with chronic conditions, were also associated in our study with a longer consultation. Although, the gender of the physician was not determinative, and neither was the gender of patient. However, other authors have found that consultations with female patients tended to be longer, especially if the physician was also female.4,5

Whether the patient was accompanied also affected consultation length. Although not included in our study, other authors also point out how increased query time is associated with age and type of ailment, possibly because of the greater complexity in ailments of the elderly.11

Pressure is an external factor to the professional (although sometimes the result of inappropriate clinical management21,28,29) which in our study was associated with a shorter consultation. It seems logical that more patients per day would mean a shorter duration. However, when communication is analysed, often such external factors are not considered, including other factors related to the patient, such as prior knowledge of them or their family.30

Lastly, all features and factors included in the multiple regression model explain 32% of the time variability in the consultation, thus leaving two-thirds unexplained. Deveugele et al.4 investigated the average consultation time in 6 European cities and the influencing factors. Factors relating to the physician and the country accounted for 22% and 23% of the variance of consultation length, whereas patient characteristics accounted for 55%.