The impact of communication is likely to be influenced by contexts, settings and patient and provider characteristics (eg. the doctor-gender preferences of patients).6 The nature of the presenting issue may also shape how communication and outcomes interact. Intimate partner violence (IPV), for example, is recognised as a sensitive issue – it is difficult to discuss and evokes embarrassment, fear, worry and shame for the patient,7 and consequently it tends to be poorly identified.
In this study, we focused on women's experiences of IPV: it affects a third of women at some point in their lives8,9 and few healthcare providers recognise it.7 Enquiry itself is challenged by numerous personal, sociocultural and professional barriers,7,10 which both reinforce and flow from the highly sensitive nature of relationship violence. Evidence on the harms and benefits of IPV identification (through universal screening programs) for women's wellbeing remains inconclusive.8 Expanding this evidence base partly requires increasing understanding about ways in which clinicians' communication and IPV disclosure interrelate. A meta-analysis of qualitative studies shows that women feel more comfortable talking about IPV when doctors communicate well.11 More quantitative evidence is needed to inform the evidence base on the specific types of communications skills required for encouraging disclosure and ensuring appropriate responses. We aimed to explore the association between specific aspects of general practitioner communication and female patients' comfort to discuss fear of a partner.
Methods
This study reports findings from the first stage of a cluster randomised controlled trial: the Women's evaluation of abuse and violence care in general practice (weave). The Human Research Ethics Committee of the University of Melbourne granted approval for the study. Methods are described in detail in the study protocol12 and findings on screening.13
Contact details of GPs representing urban/RRMA1–2 (66%) and rural/RRMA3–5 (33%) practices in Victoria were obtained from the Australian Medical Publishing Company. We excluded duplicate practices to avoid contamination, recruiting one GP per practice. General practitioners were eligible if they worked three or more sessions per week at computerised practices and had a minimum of 70% English speaking patients. Electronic lists of women meeting the inclusion criteria (age 16–50 years, consulted GP in past 12 months) were generated by the research staff. Women were excluded from the study if their GP anticipated they might encounter difficulties participating due to health issues, cognitive impairment or poor English language skills.
Eligible women were mailed a survey to assess experience of eight health/lifestyle issues in the past 12 months. The survey was adapted from the Case finding and Help Assessment Tool (CHAT), which has been validated in primary care populations.14
One item, sensitive and specific for IPV,9 asked about fear of a partner and was used to screen women for entry into the trial. The outcome variable of interest was women's comfort to discuss fear of a partner (Table 1) using a 5-point Likert scale ('very comfortable' to 'very uncomfortable'). The exposure variable of interest was a set of items measuring eight communication skills (General Practice Assessment Questionnaire [GPAQ] version 2.1)15 with responses on a 6-point Likert scale ('very poor' to 'excellent'): 'How thoroughly the GP asks about your symptoms and how you are feeling?' 'How well the GP listens to what you have to say?' 'How well the GP puts you at ease during your physical examination?' 'How much the GP involves you in decisions about your care?' 'How well the GP explains your problems or any treatment that you need?' 'The amount of time your GP spends with you?' 'The GP's patience with your questions or worries?' 'The GP's caring and concern for you?'
Table 1. Items used to assess comfort to discuss13
How comfortable would you feel talking to the GP who sent you this survey about any of the following:
- If you wanted to cut down on your smoking?
- If you were feeling down, depressed or hopeless?
- If you were afraid of your partner or ex-partner?
- If you wanted to talk about controlling what or how much you eat?
- If you wanted to cut down on your alcohol intake?
- If you needed more physical activity to be healthy?
- If you were worrying a lot about everyday problems?
- If you wanted to cut down on your drug use?
|
Statistical analysis
Analyses were performed using Stata. A total communication skills score was obtained by converting ratings to a scale score of 0 to 100. For purposes of analyses, communication items were also converted into binary variables: low (very poor to fair) versus high (good to excellent). Similarly, comfort to discuss was converted into low (very uncomfortable to neutral) versus high (comfortable to very comfortable). Participants were included in analyses regardless of whether they stated they experienced fear or not, as under-reporting is common8 and we considered the views of women who feared a partner in the past (>12 months) as also valid to our enquiry.13 Women's age, education, English as first language, actual experience of fearing a partner, GP gender and practice location were identified as potential confounders a priori.16,17 Logistic regression modelling was performed with odds ratios as the parameter estimate for analysis. Ninety-five percent confidence intervals and p values were used to assess the likely size of the association. Multivariate regression modelling examined whether the exposure outcome association was affected by confounding factors. Stepwise forward selection modelling was performed, indicating the highest log pseudo likelihood value to determine the model with the highest probability of fit to the data. Robust standard errors were used to adjust for GP clustering.
Results
Forty out of a possible 730 (5.5%) eligible GPs were recruited, including 25 female GPs and 13 GPs based rurally. Surveys were sent to 14 031 patients (averaging 350 per GP) with a response rate of 31.8% (n=4467). The mean age of patient participants was 37.3 years (SD=9.2); 79% of respondents stated that they were in a current relationship. Table 2 compares characteristics of the sample with the general population; there were no observable differences in education or marital status.
Table 2. Women's sociodemographic characteristics (N=4467)
Characteristic | weave sample* N (%) < | Australian female population** (%) |
Age (years) |
16–24 |
584 (13) |
6.8 |
25–34 |
1070 (24) |
7.1 |
35–44 |
1696 (38) |
7.1 |
45–50 |
1117 (25) |
3.5 |
English as first language |
4273 (96) |
84 |
Year 12 completed |
3080 (69) |
69.8 |
Marital status |
Married |
2298 (52) |
49# |
Divorced/separated/widowed |
543 (12) |
13 |
Never married |
1595 (36) |
38 |
Fear of partner or ex-partner† |
458 (10) |
– |
* Denominators vary due to missing data ** Australian Bureau of Statistics data # Australian women aged 15–55 years † Felt fearful a little, some, most or all the time in the past 12 months |
The mean rating for communication skills across GPs was 80.5 (SD=7, range 64.1–95.1). Female GPs were perceived to have stronger communication skills than male GPs (p<0.001). Urban GPs had lower ratings than rural GPs for the perception of the amount of time spent with their patients (p<0.001). All eight communication skills had strong associations with women's comfort to discuss fear of a partner with the GP (Table 3). In multivariate analyses, four of the 8 communication skills remained statistically significant as did the patient being older, not experiencing fear of partner in past 12 months, English as first language and lower education. Female GP was no longer associated with comfort to discuss fear. Stepwise forward selection modelling resulted in a model that consisted of all the variables used in this analysis (Table 4). The highest odds of comfort discussing fear is when patients perceive GP care and concern, followed by feeling at ease during a physical examination, involvement in decisions and time.
Table 3. Comfort to discuss fear of partner for bivariate logistic regression and multiple logistic regressions (N=4467)
| Single regression | Multivariate regression |
Variable |
Odds ratio (95% CI) |
p value |
Odds ratio (95% CI) |
p value |
Thoroughness How thoroughly the GP asks about symptoms and feelings |
5.2 (4.2–6.5) |
<0.001 |
1.0 (0.6–1.6) |
0.9 |
Listening How well the GP listens to what you say |
6.7 (5.2–8.7) |
<0.001 |
1.3 (0.8–2.2) |
0.3 |
Ease in a physical exam How well the GP puts you at ease during a physical examination |
7.9 (5.8–10.8) |
<0.001 |
1.8 (1.3–2.6) |
0.001* |
Involvement in decisions How much the GP involves you in decisions about care |
6.8 (5.4–8.7) |
<0.001 |
1.6 (1.1–2.2) |
0.007* |
Explaining How well the GP explains problems or treatment |
6.1 (4.5–8.3) |
<0.001 |
1.3 (0.8–2.1) |
0.3 |
Time The amount of time the GP spends with you |
3.9 (3.2–4.9) |
<0.001 |
1.5 (1.1–1.9) |
0.003* |
Patience The GP's patience with questions or worries |
5.9 (4.7–7.5) |
<0.001 |
1.1 (0.7–1.7) |
0.8 |
Caring The GP's caring and concern |
7.6 (6.1–9.5) |
<0.001 |
2.7 (1.9 – 4.0) |
<0.001* |
Female GP |
1.4 (1.1–1.8) |
0.006 |
1.2 (0.9–1.5) |
0.3 |
Rural practice |
1.1 (0.9–1.4) |
0.308 |
1.0 (0.8–1.3) |
0.9 |
English as first language |
1.5 (1.2–1.8) |
<0.001 |
1.5 (1.1–1.9) |
0.003** |
Fearful of a partner |
0.8 (0.6–0.9) |
0.008 |
0.7 (0.6–0.9) |
0.001** |
Age of woman (years) |
16–24 |
Base category |
|
Base category |
|
25–34 |
1.8 (1.4–2.3) |
<0.001 |
1.7 (1.4–2.2) |
<0.001** |
35–44 |
2.4 (2–2.9) |
<0.001 |
2.3 (1.9–2.8) |
<0.001** |
45–50 |
2.8 (2.2–3.4) |
<0.001 |
2.5 (2.0–3.1) |
<0.001** |
Year 12 completed |
0.7 (0.6–0.8) |
<0.001 |
0.8 (0.7–1.0) |
0.015** |
* Communication skill factors associated with the outcome (p<0.05) ** Noncommunication skill factors associated with the outcome (p<0.05) |
Discussion
Women in our study felt comfortable disclosing fear of a partner to doctors with good communication skills. Several aspects in a GP's communication influenced comfort to discuss fear of partner: spending time with the patient may influence comfort to discuss fear through simply facilitating unhurried communication;18 demonstrating care and involving patients in decisions about their care are key elements in a patient centred approach;2 putting patients at ease during physical examinations may be effective in fostering disclosure as it secures confidence in the doctor,19 building trust in the patient-doctor relationship. Consistent with research on patient centred communication,6 women in this study rated female GPs as having better communication skills than male GPs, and women were more comfortable talking to female GPs about fear of partner. However, when putting patients at ease during physical examinations was included in regression between GP gender and comfort, any association was lost. Therefore, although female GPs have higher rates of patient IPV disclosure than male GPs,7 receive higher ratings in communication skills6 and manage more psychosocial and female specific issues,20 our analyses show that it is the communication skills that count. Future studies need to further explore this link between gender, communication and disclosure.
Table 4. Parameter estimates of the final model
Variable | Odds ratio (95% CI) | p value |
Caring and concern* |
3.02 (2.2–4.1) |
<0.001 |
Ease in physical examination* |
2.05 (1.5–2.9) |
<0.001 |
Decisions* |
1.84 (1.38–2.45) |
<0.001 |
Time* |
1.61 (1.29–2.01) |
<0.001 |
Age 45–50 years** |
2.53 (2.03–3.17) |
<0.001 |
Age 35–44 years** |
2.33 (1.92–2.82) |
<0.001 |
Age 25–34 years** |
1.75 (1.37–2.24) |
<0.001 |
English as first language** |
1.48 (1.15–1.90) |
0.003 |
Year 12 completed** |
0.81 (0.69–0.96) |
0.015 |
Fear of partner or ex-partner** |
0.70 (0.57–0.86) |
0.001 |
* Communication skill factors included in the final regression model ** Noncommunication skill factors included in the final regression model |
General practitioner communication is just one factor influencing a woman's comfort in disclosure. Women currently experiencing fear were less comfortable with the prospect of talking than their nonfearful counterparts. We excluded women with poor English language skills (as the questionnaire was only in English). However, English as first language still emerged as positively associated with comfort to discuss fear. Therefore, the study may underestimate the full impact of not speaking English on discussing sensitive issues. This is important given immigrant and refugee women's vulnerability to abuse.21 Future research could explore the role of having a GP who speaks the patient's language on women's willingness to disclose sensitive issues. A lower education level was associated with willingness to volunteer a history of abuse.22 Yet findings from the Victorian context suggest that GPs tend toward asking higher educated women about IPV.7 Older women were more comfortable to disclose abuse compared to younger, more vulnerable women.21 Assessing younger women's fears, expectations and perceptions of their mental and physical health tends to be more challenging.23 Therefore, clinicians' communication with young people may be especially important.
The cross sectional design of this study limits any conclusions about causal relationships. Moreover, while the GPAQ has a growing body of data regarding validity, it may be limited in its ability to predict patient behaviours.24,25 General practitioner and patient response rates were low, and may have resulted in biased samples. However, the 12 month frequency for fear of partner reflects findings from a previous representative (78% response rate) general practice study.9 General practitioner communication scores were comparable to other Australian general practice studies.26 However, it is unknown whether weave GP scores reflect those of Australian GPs generally. Research on abuse and mental health conceivably attracts GPs who are more attuned to patient-doctor communication. Therefore, the net effect may be to underestimate the association between communication and comfort in disclosure. Excluding women with communication deficits represents an important limitation. Such women are often silenced in their experiences of violence in the home, tend to have the least 'voice' in society and are under-represented in research. This study advocates increasing communication competence to allow for greater disclosure of IPV in the primary care context. However, it also signals a need in research and practice to focus on marginalised groups and their unique needs when it comes to communicating about sensitive issues such as IPV.
Conclusion
Good doctor communication is not limited to those with an innate ability. Professional learning can assist with overcoming barriers to enquiry about IPV.27 This study identified four aspects of communication that may be emphasised in medical training programs and suggests fostering improved communication in clinicians may increase female patients' willingness to disclose IPV. However, we caveat this by pointing out that there is still a paucity of evidence about interventions that can be offered to women who disclose IPV in healthcare settings.8 Essentially, clinicians are encouraged to provide a supportive response28 that respects women's autonomy, validates experiences and emphasises safety and the human right to life free of violence.
Conflict of interest: none declared.