Family Practice Advance Access originally published online on February 3, 2006
Family Practice 2006 23(2):246-252; doi:10.1093/fampra/cmi110
The influence of GP and patient gender interaction on the duration of certified sickness absence
Mersey Primary Care R&D Consortium, Division of Primary Care, Whelan Building, University of Liverpool, Liverpool L69 3GB, UK
Correspondence to Dr Mark Gabbay, Director, Mersey Primary Care R&D Consortium, Division of Primary Care, Whelan Building, University of Liverpool, Liverpool L69 3GB, UK; Email: mbg{at}liverpool.ac.uk
Received 18 March 2005; Accepted 28 December 2005.
Shiels C and Gabbay M. The influence of GP and patient gender interaction on the duration of certified sickness absence. Family Practice 2006; 23: 246–252.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Background. Little research has focused upon how GP and patient gender interact to influence the outcome of consultation. In particular, no UK studies have investigated the effect of gender interaction on the duration of patients' certified sickness.
Objective. To investigate associations between the four GP–patient gender interaction categories and patient risk of intermediate or long-term work incapacity.
Methods. Design: Use of carbonized sickness certificates to collect routine sick note data over a 12-month collection period. Setting: Nine general practices in the Mersey Primary Care R&D Consortium. Subjects: A total of 3906 patients, certified sick by 67 GPs (including 45 GP principals). Main outcome measures: The effect of gender interaction was measured against two outcomes: intermediate (6–28 week) and long-term (28 weeks or over) periods of certified sickness.
Results. After univariate and multivariate analyses, it was discovered that certification of male patients by male GPs was significantly associated with increased prevalence of intermediate (6–28 week) certified sickness outcomes, compared with females certified by females (OR = 1.38 P = 0.009). This result was replicated in the subgroup of patients with mild mental disorder-related sickness absence. However, no association was demonstrated between gender interaction and long-term (
28 week) outcome, in the total patient group or within diagnostic subcategories.
Conclusion. GP and patient gender appear to have most impact upon sickness certification in the intermediate period. This period is already recognized as the optimum time for interventions to prevent onset of long-term incapacity, particularly in cases where the cause of sickness absence is reversible (as in psychological-related certified sickness absence). Further research is needed (particularly focusing upon attitudes and content of consultations) in order to shed more light on the gender differences found in this study.
Keywords. Sickness certification, primary care, consultation, gender interaction, GP-patient communication.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
The existing evidence concerning the impact of GP gender upon the consultation suggests that female GPs tend to be more patient-centred, less directive, spend longer with patients and demonstrate better interactive and expressive skills than their male colleagues.1–4 Perhaps as a consequence, satisfaction with female doctors is reported to be higher among both male and female patients.5,6
Whilst these findings suggest a relationship between the GP's gender and consultation outcome, it remains unclear how patient gender influences GP decisions, or whether this varies according to the consultation pairing gender mix.7 For example, there has been little research on the impact of both patient and GP gender upon decisions or negotiations regarding fitness for work within a consultation. Sickness certification is an important outcome of consultation, having various social, economic and individual implications.
In the UK, the GP has a contractual obligation to certify sickness absence on the basis of an assessment of a patient's capacity to conduct the duties involved in their ‘usual occupation’. The first week of sickness can be self-certified by the employee/patient, but any subsequent work incapacity must be certified by their GP. The most common sickness certificate is the prospective MED3, which will specify a period of time that the GP decides the patient should refrain from work, and the diagnostic reason for sickness absence. A retrospective sick note, the MED5, is also issued by GPs to certify a prior sickness episode or one diagnosed by another clinician. After a period of 28 weeks of incapacity, the decision relating to fitness for work is largely transferred from the GP to state-contracted medical officers. At this stage, the patient may be considered eligible for incapacity benefit (IB).
Although official guidelines recommend that GPs assess fitness for work primarily on the basis of the patient's health problem and any functional limitations resulting from the problem,8 there is evidence that a wide range of patient and clinician factors may have some influence upon certified sickness.9 Gender has been proposed as one of a number of individual patient characteristics associated with periods of work incapacity.9 Whilst research has demonstrated a tendency for women to be sick listed more often, and to have lower work resumption rates than men,10 there is no consistent evidence indicating an independent effect of patient gender on duration of certified sickness. The results of studies of the influence of clinician gender in this area are also conflicting. Weak evidence suggests female doctors are more likely to certify sickness than are their male counterparts.11,12 However, other research has found no independent relationship between clinician gender and either the length of certified sickness episodes or attitudes to certification.13
A Norwegian study examined the relationship between primary care physicians' assessments of work ability in sicklisted patients, and the gender of both parties. This reported that male doctors assessing female patients were significantly less likely to consider their work ability to be reduced, and more likely to only issue part-time certification.14
The most rigorous research in the field of sickness certification has tended to be conducted in Scandinavian countries such as Norway and Sweden. A major reason for this has been the opportunity for researchers in these countries to access large datasets containing comprehensive details of benefit claimants at all stages of incapacity. Regional and national databases of certification data have been maintained in line with the development of their social insurance systems. A major obstacle to such research in the UK has been the absence of similar data resources. Although the Department for Work and Pensions (DWP) publishes regular profiles of claimants and beneficiaries of incapacity-related benefits, there is little known about those patients who are certified in the first 28 weeks of sickness, before they become eligible for IB. During this period, because most of the financial cost of sickness absence is borne by the individual employers, there are no centralized official data relating to patients taking certified sick leave. Hence, only broad trends in sickness certification can be discerned over this period.
It has been reported that, every week in the UK, about 17 000 people of working age reach their sixth week of sickness absence. The majority will subsequently resume their paid employment, but it is estimated that for about 3000 the incapacity episode will exceed 28 weeks, when they become eligible for IB (Short-term/higher rate). Of those claiming this benefit, 
40% will remain on it for a year or more. The majority of these claimants will never return to the workforce.15
Our study aimed to
- describe differences between patient characteristics within the four GP-patient gender interaction categories in the sickness certification process (i.e. male patients certified by male or female GPs, and female patients certified by female or male GPs).
- investigate associations between these gender interaction categories and prevalence of intermediate (6–28 weeks) or long-term (28 or more weeks) certified sickness episodes.
- conduct similar analyses within the largest diagnostic subgroups of patients in the study.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Sickness certification data
In the absence of any recognized method of systematically capturing routine sickness certification data at general practice level, we collaborated with the Department of Social Security to develop and utilize carbonized copies of MED3 and MED5 certificates, enabling capture of all sick notes issued in participating practices during the sampling period. This method ensured that, after the issue of a sick note, a copy of details was retained and entered into our anonymized baseline database. All GPs in the nine practices of the Mersey Primary Care Research and Development Consortium (including 45 GP principals) agreed to use the pads of carbonized certificates for a period of 12 months (across 2000 and 2001).
The process of data collection involved an administrative staff member at each practice collating the carbonized details of sick notes and entering them into a local database. At this stage, patients were anonymized, but a study number was allocated to each one for tracking purposes. At the end of the collection period, anonymized data from all nine general practices were merged into the one large database at the consortium research centre.
The four research ethics committees (Liverpool, South Sefton, North Sefton, South Cheshire), covering the geographical area of the Consortium, approved the use of these certificates as a means of retaining routine sick note data.
Our baseline dataset includes anonymized details of 13 127 certificates issued to 6271 patients. The fields in the initial database included patient gender, age, Townsend social deprivation score (computed from postcode), reason for incapacity, duration of sick note and certifying GP code. The gender (if known) of the certifying GP was added to the data at a later stage. The reason for incapacity reported on the individual certificate was re-coded into a broader diagnostic category (based upon READ categories, refined for a general practice-based study).16 Further methodological detail and analysis of the patient data have been reported elsewhere.17
Of the original 6271 patients, 3906 met our inclusion criteria for this aspect of the study. Namely
- patients having continuous sick notes, within the same diagnostic category and issued by a single GP, throughout the data collection period.
- cases where both the patient and GP gender were known and recorded.
Outcome measures
The two main outcome variables in the study were based upon important junctures in the transition to long-term incapacity: an intermediate period of certified sickness (6 weeks up to 28 weeks) and a long-term period (28 weeks or more).
Statistical analysis
The four possible patient–GP gender interactions were the focus of analysis in the study. A new variable was created in the database, allocating each patient to one of the interaction categories. Analysis included investigation of differences across the four interaction categories, and association between the categories and certified sickness outcomes. The period of uninterrupted certified sickness in the 12-month period was totalled for each patient in the study (regardless of the number of sick notes issued to the individual patient). The total duration of certified sickness for each patient was transformed into two binary variables: (i) intermediate sickness episode (6–28 week) or <6 weeks and (ii) long-term (28 week or more) episode or <28 weeks.
At the univariate level of analysis, significant differences in continuous data across patients in the four interaction groups were investigated by application of a one-way ANOVA, with post hoc analysis (Tukey HSD test) of pairs of categories. Chi-square tests were implemented to explore associations between GP and patient gender interaction and other categorical variables (e.g. diagnostic category of sickness episode, the binary outcome variables). A continuity correction was added for 2 x 2 tables.
At the multivariate level, logistic regression models were constructed to control for various effects of patient casemix (age, social deprivation) and diagnostic category of episode, in order to investigate independent associations between GP and patient gender interaction and prevalence of intermediate and long-term incapacity. The effect of each interaction category on outcome is reported in terms of an odds ratio, confidence intervals and the relevant P-value.
In order to investigate results of the main analysis in more detail, multivariate analyses of association between GP and patient gender interaction and certified sickness outcomes were also conducted for selected subgroups of patients; those incapacitated due to each of the five most important diagnostic categories in the study.
For all univariate and multivariate analyses, a P-value of 0.05 was defined as the level of statistical significance.
Data were analysed using SPSS for Windows 11.0.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Characteristics of patients in GP and patient gender interaction categories
The dataset included sickness episodes certified by 67 GPs (including 45 principals). Of these GPs, 32 (47.8%) were male. Nearly 55% of the 3906 patients who had certified sickness episodes were female.
Male GPs certified similar proportions of male and female patients. Nearly 52% (1254/2424) of their patients were male. However, female patients accounted for 65.8% (975/1482) of episodes certified by a female GP (
2 = 114.0 d.f. = 1 P < 0.001).
Table 1 summarizes patient characteristics within the four gender interaction categories. There was a significant difference in mean age across the categories (F = 8.3 P < 0.001). Post hoc analysis (Tukey HSD) indicates that male patients certified by a male GP were significantly older than males (mean age 40.8 versus 38.4; P < 0.001) or females certified by a female GP (mean age 40.8 versus 38.6; P < 0.001).
|
In terms of social deprivation, there was a significant difference in mean Townsend deprivation scores across the four groups of patients (F = 28.6 P < 0.001). In particular, male patients seeing a female GP tended to be significantly more deprived than females (P < 0.001) or males (P < 0.001) having sickness certified by a male GP.
For all gender interaction groups, the main reason for certification (in terms of diagnostic categories) tended to be related to mild mental disorders (MMD) or musculoskeletal problems. These two diagnostic categories accounted for over half of the days certified in each GP-patient group. However there were some differences in the diagnostic profiles of the gender categories. Most striking is the greater importance of musculoskeletal problems in males seeking certification from male GPs, and the relatively higher prevalence of MMDs in women certified by female GPs.
Apart from the two main diagnostic categories, the most important certified diagnosis was respiratory-related for female patients certified as sick by male GPs (6.6% of total days certified) and those women seen by a female GP (6.1%). For those males consulting a male GP, the third most prevalent diagnostic category was ‘circulatory system disease’ (8.4%), while the reason for certification was recorded as ‘post-op recovery’ in 6% of the certified sickness absence of male patients consulting a female GP.
GP–patient gender interaction and duration of certified sickness episodes: all patients
The final two columns of Table 1 report rates of intermediate and long-term certified sickness episodes, for each of the GP-patient gender categories. Proportions of patients receiving both periods of certification differed significantly across the four GP-patient gender categories (intermediate: 2 = 20.1 d.f. = 3 P < 0.001; long-term: 2 = 21.5 d.f. = 3 P < 0.001). In terms of 6–28 week incapacity, the highest rate within a category was found in the male GP-male patient category (26.4%) and the lowest when females were certified by female GPs (18.3%). For long-term incapacity, trends were similar, with male–male and female–female categories having rates of 9.2 and 4.8%, respectively.
Logistic regression models were constructed in order to investigate independent effects of gender interaction on outcomes, controlling for patient age and social deprivation score, and also for diagnostic category of the sickness episode. The female GP-female patient category, with the lowest prevalence of intermediate and long-term certified sickness outcomes in univariate analysis, was classed as the reference criterion in the multivariate models. Table 2 presents results of regression analysis using the intermediate outcome as the dependent variable. After controlling for the patient-based and diagnostic factors, the likelihood of an intermediate sickness episode (6–28 week) was significantly raised for male patients certified by male GPs, compared with the reference category (OR = 1.38 95% CI 1.09–1.75). (When the male GP-male patient category was used as the reference, no significant difference was found between it and either of the two categories, ‘male–female’ and ‘female–male’, in relation to outcome.)
|
Similar multivariate analysis found no significant effect of gender interaction on the long-term (28 week or more) outcome, after controlling for the same set of patient-based and diagnostic covariates. (Not reported in table form.)
GP–patient gender interaction and duration of certified sickness episodes: selected patient diagnostic subgroups
In terms of contributing to the total days certified across all patients in the 12-month study period, the five most important diagnostic categories were ‘MMD’ (39.7% of sick days certified), ‘Musculoskeletal’ (17.03%), ‘Injury/poisoning’ (6.1%), ‘Circulatory disease’ (5.8%) and ‘Respiratory disease’ (4.9%).
Logistic regression models were constructed for each of these five subgroups of patients in order to focus upon the association between GP and patient gender interaction and both intermediate and long-term incapacity, controlling for patient age and deprivation score. In order to maintain consistency, the female patient–female GP interaction was the reference category in each model.
In terms of association with intermediate (6–28 week) outcome, the results of the regression analyses reported in Table 3 indicate that for only one diagnostic subgroup of patients, those suffering from MMD, is there a gender interaction effect similar to that in the main analysis. Male patients certified for this type of diagnosis by a male GP were 1.64 times more likely than female equivalents certified by a female to be incapacitated for 6–28 weeks (95% CI 1.13–2.38 P = 0.009).
|
For all diagnostic groups, including the MMD category of patients, there were no significant associations between GP and patient gender interaction and long-term (28 or more) certified sickness outcomes, after adjustment for patient age and deprivation. (Not reported in table form.)
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
The strengths and weaknesses of this study are similar to those reported in previous published research based upon the same dataset.17 Its major strength is its innovative nature. This is the first UK-based study of the influence of GP and patient gender interaction on the process of sickness certification. Our baseline dataset is the largest of its type collected in the UK to date. The most important methodological weakness is the absence in the certification database of key patient variables. Although patient (and GP) gender was easy to establish, other important patient data (e.g. relating to occupation and family circumstances), with potential implications for certified sickness, were not able to be collected for the study. A more study-specific limitation was the lack of prior information held about patients' preference for a GP of a certain gender or if some GPs in the study had disproportionate numbers of male or female patients on their list.
Our general findings suggest that men consulting male GPs were significantly more likely to be certified for an intermediate (6–28 weeks) period, even when controlling for other patient factors associated with such decisions (age, diagnosis and deprivation). However, we were not able to demonstrate a significant influence for consultation gender interaction on the development of long-term incapacity. There appears to be a threshold (28 weeks of incapacity) when gender interaction no longer has a significant independent association with work incapacity.
What is not clear is whether these differences are linked to the relative health of those presenting to their GP, the content of the interaction during the consultation, or GP factors and attitudes. It may be that assumptions about roles within work for male and female patients, and hence capacity for work, may differ between GPs of different gender. There is also the possibility that male GPs assume that their male patients in this situation have relatively greater problems before taking a decision to consult, or have a relatively reduced capacity to work, or slower rehabilitation after illness. On the other hand, the key to gender interaction differences might be found with the patient. Male patients may be more demanding, or better negotiators, when facing a male GP. What is not clear is whether this group do indeed have relatively greater problems, poorer coping skills, or are more sympathetically dealt with by male than female GPs.
Our results specifically suggest that this difference in outcome of gender interaction is most pronounced when the patient is certified sick after presenting with a MMD, such as depression, anxiety or stress. The relatively increased likelihood of male patients receiving longer periods of certification from male GPs may be related to evidence indicating particular difficulties in male patients' presentation of psychological symptoms and the subsequent creation of barriers to GP detection and a definitive diagnosis of an MMD.18 In this situation, where diagnosis may be relatively complex, longer certified sickness absence may have a role to play in negotiation between GP and patient. The male patient might be willing to accept the recording of ‘depression’ on a sick note, but be reluctant to accept this as a diagnosis that requires his compliance with a particular treatment regimen. While this is one possible explanation for gender differences in MMD-related certified sickness, the gender of the certifying GP may accentuate the difference. Research studies have indicated that, compared with their male colleagues, female doctors tend to have interpersonal skills more attuned to dealing with psychological problems presented by patients. They are seen as engaging in more ‘patient-centred’ communication,7 managing more problems of a psychosocial nature, and being more effective in psychosocial question-asking and counselling.2 Lack of these attributes in the male GP may contribute to the delay in resolving male psychological problems and returning the patient to work.
These assumed interpersonal skills of the female GP may also partly explain why rates of intermediate-term MMD-related certified sickness are lowest among female patients certified by female GPs. The communication skills reported above may help the female GP in establishing an early diagnosis in female patients, who themselves may be more forthcoming in presenting symptoms of psychological distress than their male counterparts. The female GP may also better understand the implications (financial and otherwise) of prolonged work incapacity for a patient of her same gender, and in turn, the female patient may be more willing to accept advice relating to work resumption from a female doctor. There is also the possibility that the female GP may use her own gender experience (of working, family circumstances etc.) in a judgement of how incapacitated a female patient really is. This judgement may lead to advice on an early return to work, which could in itself be therapeutic for the patient.
Given the scarcity of the evidence in the sickness certification literature, the reasons offered for gender interaction differences are necessarily speculative. In any case, if valid, they are only relevant to the intermediate stages of certified sickness. Factors such as different patient presenting styles and differential GP communication skills do not seem to make much difference to long-term (28 week or more) patient incapacity for work. However, this intermediate (6–28 week) period of sickness absence is seen as a crucial juncture in designing and applying interventions to prevent onset of long-term incapacity,19 particularly when the causes of incapacity are potentially reversible (as in most MMDs).Therefore, an understanding of how patient and GP gender interact to influence certified sickness in this period provides a useful contribution to the wider task of identifying the range of significant factors in the route to long-term work incapacity.
In a context of a dearth of research in this area, this quantitative study can be considered a baseline attempt to establish if there is an association between certified sickness absence and the gender of the agencies participating in the certification process. Further quantitative research, preferably using more patient-based covariates, is required to confirm the findings and explore explanatory hypotheses. Contextually based qualitative studies would have to be conducted in order to make more sense of the broad survey findings. In particular, observation of consultations involving different gender pairings may help to identify aspects of the interaction between the GP and patient that lead to a specific sickness certification outcome. Research could usefully explore attitudes (to certified sickness and wider related issues) that both parties bring to the consultation, and how gender influences these, within social and occupational contexts.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Funding: This project was funded by Mersey Primary Care R&D Consortium (NHS R&D Support Funding).
Ethical approval: Liverpool, North Sefton, South Sefton, South Cheshire LRECs (2000).
Conflicts of interest: none.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
1 West C. Reconceptualizing gender in physician-patient relationships. Soc Sci Med 1993; 36: 57–66.[CrossRef][Web of Science][Medline]
2 Hall JA, Irish JT, Roter DL et al. Gender in medical encounters: an analysis of physician and patient communication in a primary care setting. Health Psychol 1994; 13: 384–392.[CrossRef][Web of Science][Medline]
3 Law SA, Britten N. Factors that influence the patient centredness of a consultation. Br J Gen Pract 1995; 45: 520–524.[Web of Science][Medline]
4 Roter DL, Hall JA, Aoki Y. Physician gender effects in medical communication: a meta-analytic review. JAMA 2002; 288: 756–764.
5 Bertakis KD, Helms LJ, Callahan EJ et al. The influence of gender on physician practice style. Med Care 1995; 33: 407–416.[CrossRef][Web of Science][Medline]
6 Delgado A, Lopez-Fernandez LA, de Dios Luna J. Influence of the doctor's gender in the satisfaction of the users. Med Care 1993; 31: 795–800.[Web of Science][Medline]
7 Zaharias G, Piterman L, Liddell M. Doctors and patients: gender interaction in the consultation. Acad Med 2004; 79: 148–155.[CrossRef][Web of Science][Medline]
8 Department for Work and Pensions. Guide for Registered Medical Practitioners. London: DWP; 2004.
9 Alexanderson K. Sickness absence: a review of performed studies with focus on levels of exposures and theories utilised. Scand J Soc Med 1998; 26: 241–249.[Web of Science][Medline]
10 Bloch FS, Prins R. Who Returns to Work and Why? London: Transaction; 2001.
11 Englund L, Tibblin G, Svärdsudd K. Variations in sick-listing practice among male and female physicians of different specialties based on case vignettes. Scand J Prim Health Care 2000; 18: 48–52.[CrossRef][Web of Science][Medline]
12 Englund L, Svärdsudd K. Sick-listing habits among general practitioners in a Swedish county. Scand J Prim Health Care 2000; 18: 81–86.[CrossRef][Web of Science][Medline]
13 Tellnes G, Sandvik L, Moum T. Inter-doctor variation in sickness certification. Scand J Prim Health Care 1990; 8: 45–52.[Medline]
14 Brage S, Reiso H. Work ability and gender-physicians' assessment of sicklisted patients. Tidsskr Nor Loegeforen 1999; 25: 3737–3740.
15 Waddell G, Burton AK, Main CJ. Screening to Identify People at Risk of Long-term Incapacity for Work. London: RSM; 2003.
16 O'Neil M, Payne C, Read J. READ codes Version 3; a user-led terminology. Methods Inform Med 1995; 34: 187–192.[Web of Science][Medline]
17 Shiels C, Gabbay M, Ford F. Patient factors associated with duration of certified sickness absence and transition to long-term incapacity. Br J Gen Pract 2004; 54: 86–91.[Web of Science][Medline]
18 Shiels C, Gabbay MB, Dowrick C, Hulbert C. Depression in men attending a rural general practice: factors associated with prevalence of depressive symptoms and diagnosis. Br J Psychiatry 2004; 185: 239–244.
19 Waddell G, Burton AK. Concepts of Rehabilitation for the Management of Common Health Problems. London: TSO; 2004.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G Wynne-Jones, C. Mallen, C. Main, and K. Dunn Sickness certification and the GP: what really happens in practice? Fam. Pract., December 24, 2009; (2009) cmp096v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||