Abstract
Objectives
To investigate the effect of physician gender on consultation length in UK hospital outpatient clinics and compare this, through meta-analysis, with previous studies outside the UK.
Design
Observational data on clinic times were analysed and findings were combined in a meta-analysis with existing studies investigating the effect of physician gender on consultation length.
Setting
UK hospital practice.
Participants
A total of 174 observations of outpatient consultations with 10 hospital specialists (consultants) from different specialties in two UK hospital trusts.
Main outcome measures
Clinic times were recorded and analysis of consultation length was undertaken with physician gender as a covariate. Data were then synthesised through meta-analysis with 10 existing studies in this field.
Results
No statistically significant difference was found in the length of consultations for male and female doctors in these UK hospital settings. When pooled with existing studies, consultations with women doctors were found to be approximately two minutes longer than with men (p = 0.01).
Conclusions
Findings from this analysis of clinic consultations in the UK National Health Service do not support previous studies, which were undertaken predominantly in North America and primary care settings. Overall, meta-analysis suggests doctors’ gender may influence consultation length. Gender differences in communication should be considered in training clinicians and in overall clinical practice.
Keywords: physician gender, physician–patient communication
Introduction
Healthcare providers and policy makers are increasingly concerned with securing improvements in productivity of the healthcare workforce, in an international climate of fiscal austerity and increasing demand for healthcare.1 Gender differences in the activity rates of hospital consultants in the UK2 and Canada3 highlight the need for research that explores potential sources of variations in doctors’ productivity, particularly given increasing numbers of women entering the medical workforce.
Several studies have been undertaken to explore the effect of doctors’ gender on communication that takes place during medical consultations, with findings suggesting that female doctors engage in a range of communication styles that may be associated with longer consultation times, including a more ‘partnership building’ approach which encompasses behaviours such as encouragement, reassurance, lowered dominance and positive talk.4–12 In a recent systematic review, the effect of doctors’ gender on consultation times was explored using meta-analysis to pool results across 10 studies in this field. Results indicated that female doctors spend over two minutes longer per consultation compared to male doctors.4 These findings are based largely on studies in settings outside the UK (9/10) and in primary care (7/10) and, to the authors’ knowledge, there is currently no research exploring the effect of gender on consultation length in a UK hospital setting.
This study aimed to address this gap in the literature by analysing the effect of doctors’ gender on consultation length using observations undertaken as part of a larger qualitative exploration of the working lives of hospital consultants in England. These findings are synthesised with findings from existing studies4 using meta-analysis.
Methods
The study was carried out in two stages, first analysing gender differences in hospital outpatient clinic consultation times and then pooling findings with existing studies to obtain a combined estimate of the effect of physician gender on consultation length.
Analysis of observed time data from clinic consultations
Consultation length data were collected as part of a qualitative observational study which explored gender differences in hospital consultants’ working lives, including interactions with patients. Consultation length was recorded by non-participant observers using wristwatches during outpatient clinic consultations with 10 hospital consultants (four men and six women) in two National Health Service (NHS) hospital trusts in the north of England. A range of both medical and surgical specialties were included (general medicine, general surgery, oncology, ophthalmology and palliative care). Owing to the small number of consultants in each specialty these were grouped into the following categories for the purpose of summarising the data: medicine, surgery or oncology. A separate category for oncology was deemed necessary as these consultations were considerably longer in length.
NHS research ethics committee approval (10/H0401/76) and university departmental research ethics committee approval were obtained for this study.
Descriptive statistics were used to explore the mean consultation length for male and female consultants and by specialty. Consultation length by consultant was also described in order to explore the variation across the sample. Robust standard errors (Huber-White sandwich estimator) regression models were used to explore the effect of doctors’ gender on consultation length as implemented in Stata version 13 (StataCorp, College Station, TX, USA). The individual consultation was used as the unit of analysis but the standard errors were adjusted to account for the potential clustering by consultant. Standard statistical tests make the assumption that individuals are independent of one another. In this analysis, the data were naturally ordered by consultant (i.e. not independent of one another), hence this needed to be taken into account in the analysis.13,14
Synthesising observational data with existing studies
Findings from this analysis were then synthesised with results from previous studies in this field. A systematic review4 identified 10 existing studies5–7,11,15–20 from which similar data were available. This review assessed the effect of physician gender on communication during medical consultations, including various domains such as psychosocial communication style and consultation length. Seven electronic databases were searched with no date or language restrictions (MEDLINE, PsychINFO, EMBASE, CINAHL, Health Management Information Consortium, Web of Science and ASSIA). Studies were included that presented interpretable primary data; studied physicians and actual patients; measured communication using neutral observers, audiotape or videotape; and tested for an association between physicians’ gender and at least one interpretable communication variable. Studies using physician reported length of consultation were excluded, as were studies of psychiatric consultations or those involving medical students and simulated patients. Therefore, the observational time data, collected in stage one of this study, were eligible for combining with these existing studies. Database search strategies can be found in Appendix 1 (available online as Supplementary material) and full details of this search and search results can be found in the original article.4
The results from the analysis of observational clinic data on consultation length from stage one of this study were synthesised with these existing study results using a random effects meta-analysis. The degree of heterogeneity across studies was assessed using χ2 and I2 tests.21 All analyses were conducted in Stata version 13 (StataCorp, College Station, TX, USA), using the metan command.
Results
Consultation length in UK hospital outpatient clinics
In total, 174 patient consultation times were recorded (63 with male doctors and 111 with female doctors). Hospital outpatient consultations with male doctors in this study were slightly longer than for female consultants, although this difference was small (mean [SD]: 12.51 mins [6.09] and 11.86 mins [6.01], respectively) (Table 1). Visit length varied by specialty, with oncologists holding the longest consultations and surgeons the shortest (Table 1). The variability in consultation length across participants is demonstrated in Table 2.
Table 1.
Visit length (in minutes) for men, women, all consultants and model findings.
| Mean (SD) | Range (min–max) | Observations (N) | Difference in consultation length | 95% Confidence intervals | p value | |
|---|---|---|---|---|---|---|
| Gender | ||||||
| Men | 12.51 (6.09) | 3–26 | 63 | – | – | – |
| Women | 11.86 (6.01) | 3–35 | 111 | −0.64 | −4.41 to 3.13 | 0.71 |
| Specialty | ||||||
| Medicine | 12.14 (5.59) | 4–25 | 43 | |||
| Surgery | 11.42 (5.62) | 3–34 | 106 | |||
| Oncology | 14.88 (7.69) | 3–35 | 25 | |||
| Total | 12.10 (6.03) | 3–35 | 174 | – | – | – |
Table 2.
Number of consultations recorded and mean consultation length by consultant.
| Consultant | Gender | Specialty | Total observations (N) | Clinics (N) | Visit length (mean, SD) | Visit length (min–max) |
|---|---|---|---|---|---|---|
| FMedA1 | Female | Medicine | 12 | 1 | 10.92 (5.52) | 5–25 |
| FMedA4 | Female | Medicine | 19 | 2 | 13.84 (4.98) | 5–24 |
| MMedA5 | Male | Medicine | 12 | 1 | 10.67 (6.27) | 4–23 |
| FSurgA6 | Female | Surgery | 25 | 2 | 10.76 (4.52) | 3–20 |
| FSurgA7 | Female | Surgery | 26 | 2 | 9.96 (5.97) | 4–27 |
| MSurgA8 | Male | Surgery | 26 | 2 | 14.31 (4.62) | 8–24 |
| FSurgB9 | Female | Surgery | 17 | 2 | 13.12 (7.04) | 4–34 |
| MSurgB10 | Male | Surgery | 12 | 1 | 7.33 (2.71) | 3–12 |
| FMedB11 | Female | Oncology | 12 | 4 | 14.33 (8.05) | 4–35 |
| MMedB12 | Male | Oncology | 13 | 3 | 15.38 (7.64) | 3–26 |
When including gender in a regression model to assess the relationship between these variables and length of clinic consultations, gender does not appear to be a predictor of consultation length in these observations (p = 0.71) (Table 1).
Synthesis with existing studies
When combined with the results from existing studies, a statistically significant difference in the length of consultation for male and female doctors remained (coefficient = 1.95; 95% CIs 0.44 to 3.46 and P = 0.01). The forest plot (Figure 1) provides a graphical representation of this, with the findings from this study labelled ‘Jefferson et al. 2014’. Statistical tests for heterogeneity revealed significant variation across studies (χ2 = 30.17, df = 10, p = 0.001, I2 = 66.9%).
Figure 1.
Forest plot of consultation length.
Discussion
This is the first study that has measured gender differences in the length of hospital consultants’ outpatient clinic consultations in a UK setting. Although the sample size of this study was small as it formed part of an in-depth qualitative study, the findings demonstrate a need for further research in this field as they are not entirely consistent with existing evidence. This may be of particular benefit as this study took place in a different setting to many others in this field.
When synthesising the observational data collected here with existing studies in this field, a slightly smaller estimate of effect is generated to that which has previously been found,4 owing to the lack of statistically significant gender difference found in the present study. Nevertheless, this pooled estimate remains statistically significant and may be a potentially important difference at almost two minutes longer per consultation with female doctors compared with male doctors. This difference may have a large impact over the course of a doctor’s overall working day and may partly explain the gender differences in activity rates of hospital consultants which have previously been reported.2,3 Based on UK workload data for General Practitioners (GPs) (working an average of 38.2 h per week and spending 11.5 min per consultation),22 an additional two minutes per consultation would result in female GPs seeing approximately 15% fewer patients than male GPs during the course of a working day. Alternatively, this difference could lead to female doctors spending longer at work, as Roter et al.23 suggest, a difference of two minutes per consultation could result in an extra hour of work at the end of a busy day.
The results of this meta-analysis should be interpreted with caution, as the previous review of the quality of existing studies in this field has shown that these studies are generally small and tend to lack rigorous research methods.4 For example, the 10 existing studies in this field have a median sample size of only 27 doctors (IQR: 17 to 49.5) and, given the small sample size, are often unable to control for important confounding variables. High variability in study findings and heterogeneity is also a problem, as demonstrated in Figure 1. This may relate to the variability in study methods, practice settings, patient groups, and doctor characteristics, such as years of experience, in these studies.
Strengths and limitations
The main limitation of this study is the small sample size of doctors as, although these findings are based on 174 observations of consultation length, they may not be generalisable to wider samples of doctors. Further, this study was not powered to detect a prespecified effect size and hence, there may have been too few doctors measured in order to detect a statistically significant difference should one exist. Small sample sizes are also a common problem of existing research in this field and this limits the ability to control for potential confounders, such as specialty, which was not appropriate in this analysis due to the small number of clusters.
We chose to adjust for the potential clustering within consultant using robust standard errors. There are a number of different ways in which the analysis could have been performed, for example using cluster summary statistics, general estimating equations, or multilevel modelling. We chose to use robust standard errors as we were interested in a cluster level covariate. Given the small number of consultants included in the analysis we checked the robustness of the findings using a bootstrap-based improvement24 and the results were comparable. We also checked for consistency of results using model-based approaches to the analysis, with and without corrections for small sample sizes.25 All models produced estimates in the same direction but the magnitude of effect was different depending on which analysis was undertaken. Further research is required to explore the most appropriate method of analysis when exploring cluster level covariates with a small number of clusters.
A further potential weakness of this study may be the method of measuring time, which was done with wrist watches rather than stopwatches, which would provide more accurate measurement. The use of stopwatches in the present study was considered too intrusive as they could have potentially altered consultants’ behaviours and interactions with patients.
Conclusions
This study reports gender differences in clinic consultation lengths for the first time in a UK hospital setting. Further research is warranted in a larger sample of UK hospital doctors. In future studies it is important to consider adjusting for potential confounding variables when measuring gender differences in length of consultations such as specialty and patient gender.26
In the UK hospital setting, healthcare organisations may wish to consider other potential sources of variability in consultation times, as there was a wide range of times recorded during observations. For example, the streamlining of processes that surround the medical consultation, such as note making or completion of diagnostic tests which appeared to influence the total time spent per patient in observations in this study, may result in variations in consultation length.
Although no gender difference was found in hospital consultants’ length of clinic consultations in this UK setting, when considering the evidence base of studies internationally and in different settings, an overall gender effect does appear to exist, with women spending longer on patient consultations compared to men. This has implications for healthcare organisations internationally in terms of the doctors’ productivity, particularly as the gender composition of the medical workforce rapidly approaches parity.
Declarations
Competing interests
None declared
Funding
Karen Bloor and Laura Jefferson were funded by a Career Development Fellowship from the National Institute of Health Research in England. The views expressed in this paper are those of the authors and not necessarily those of the NHS, The NIHR or the Department of Health.
Ethical approval
NHS research ethics committee approval (10/H0401/76) and university departmental research ethics committee approval were obtained for this study.
Guarantor
Laura Jefferson is study guarantor. All authors had access to the data in the study and can take responsibility for the integrity and accuracy of the data analysis.
Contributorship
LJ and KB initiated the idea for the study and undertook data collection. LJ and CH undertook data analysis and interpreted the results with discussion with KB. LJ wrote the initial draft of the paper, to which all the authors contributed.
Acknowledgements
We are grateful to the hospital doctors who participated in this research.
Provenance
Not commissioned; peer-reviewed by Lindsay Hedden
References
- 1.Department of Health. NHS Chief Executive’s Annual Report 2008/2009, London: Department of Health, 2009. [Google Scholar]
- 2.Bloor K, Freemantle N, Maynard A. Gender and variation in activity rates of hospital consultants. J R Soc Med 2008; 101: 27–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Woodward CA, Hurley J. Comparison of activity level and service intensity of male and female physicians in five fields of medicine in Ontario. CMAJ 1995; 153: 1097–1106. [PMC free article] [PubMed] [Google Scholar]
- 4.Jefferson L, Bloor K, Birks Y, Hewitt C, Bland M. Effect of physicians’ gender on communication and consultation length: a systematic review and meta-analysis. J Health Serv Res Policy 2013; 18: 242–248. [DOI] [PubMed] [Google Scholar]
- 5.Bernzweig J, Takayama JI, Phibbs C, Lewis C, Pantell RH. Gender differences in physician-patient communication. Evidence from pediatric visits. Arch Pediatr Adolesc Med 1997; 151: 586–591. [DOI] [PubMed] [Google Scholar]
- 6.Hall JA, Irish JT, Roter DL, Ehrlich CM. Gender in medical encounters: an analysis of physician and patient communication in a primary care setting. Health Psychol 1994; 13: 384–392. [DOI] [PubMed] [Google Scholar]
- 7.Roter D, Lipkin M, Jr, Korsgaard A. Sex differences in patients’ and physicians’ communication during primary care medical visits. Med Care 1991; 29: 1083–1093. [DOI] [PubMed] [Google Scholar]
- 8.Shapiro J. Correlates of family-oriented physician communications. Fam Pract 1999; 16: 294–300. [DOI] [PubMed] [Google Scholar]
- 9.Wolfensberger JA. The impact of gender, gender attitudes, and communication on patient satisfaction during medical visits. Dissertation Abstr Inter: Sec B: Sci Eng 1997. [Google Scholar]
- 10.Zandbelt LC, Smets EM, Oort FJ, Godfried MH, de Haes HCJ. Determinants of physicians’ patient-centred behaviour in the medical specialist encounter. Soc Sci Med 2006; 63: 899–910. [DOI] [PubMed] [Google Scholar]
- 11.Meeuwesen L, Schaap C, van der Staak C. Verbal analysis of doctor-patient communication. Soc Sci Med 1991; 32: 1143–1150. [DOI] [PubMed] [Google Scholar]
- 12.Ainsworth-Vaughn N. Topic transitions in physician-patient interviews: power, gender, and discourse change. Lang Soc 1992; 21: 409–426. [Google Scholar]
- 13.Bland JM, Kerry SM. Statistics notes. Trials randomised in clusters. BMJ 1997; 315: 600–600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Donner A, Klar N. Design and Analysis of Cluster Randomization Trials in Health Research, London: Arnold Publishers Limited, 2000. [Google Scholar]
- 15.Bertakis KD, Franks P, Azari R. Effects of physician gender on patient satisfaction. J Am Med Wom Assoc 2003; 58: 69–75. [PubMed] [Google Scholar]
- 16.Bertakis KD, Helms LJ, Callahan EJ, Azari R, Robbins JA. The influence of gender on physician practice style. Med Care 1995; 33: 407–416. [DOI] [PubMed] [Google Scholar]
- 17.Carr-Hill R, Jenkins-Clarke S, Dixon P, Pringle M. Do minutes count? Consultation lengths in general practice… including commentary by Howie J. J Health Serv Res Policy 1998; 3: 207–214. [DOI] [PubMed] [Google Scholar]
- 18.Pahal JS, Li HZ. The dynamics of resident-patient communication: data from Canada. Comm Med 2006; 3: 161–170. [DOI] [PubMed] [Google Scholar]
- 19.Roter DL, Geller G, Bernhardt BA, Larson SM, Doksum T. Effects of obstetrician gender on communication and patient satisfaction. Obstet Gynecol 1999; 93: 635–641. [DOI] [PubMed] [Google Scholar]
- 20.van Dulmen AM, Bensing JM. Gender differences in gynecologist communication. Women Health 2000; 30: 49–61. [DOI] [PubMed] [Google Scholar]
- 21.Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.NHS Information Centre for Health and Social Care. GP workload survey, London: The Information Centre, Primary Care Statistics, 2007. [Google Scholar]
- 23.Roter DL, Hall JA, Aoki Y. Physician gender effects in medical communication: a meta-analytic review. JAMA 2002; 288: 756–764. [DOI] [PubMed] [Google Scholar]
- 24.Cameron A, Gelbach J, Miller D. Bootstrap-based improvements for inference with clustered errors. Rev Econ Stat 2008; 90: 414–427. [Google Scholar]
- 25.Kenward M, Roger J. Small sample inference for fixed effects from restricted maximum likelihood. Biometrics 1997; 53: 983–997. [PubMed] [Google Scholar]
- 26.Sandhu H, Adams A, Singleton L, Clark-Carter D, Kidd J. The impact of gender dyads on doctor-patient communication: a systematic review. Patient Educ Couns 2009; 76: 348–355. [DOI] [PubMed] [Google Scholar]