Sexual identity and utilisation of primary healthcare services: findings from the New Zealand Health Survey

Access to primary healthcare services is paramount in achieving the best possible health outcomes.¹ Failure to engage with healthcare services may lead to unmet health needs, preventable hospitalisation and serious illness. In many instances, the prevention of serious ill health is contingent on early detection of problems to enable effective intervention. Quality of care is also important because it can be a determinant of ongoing engagement with healthcare professionals. Despite this, engagement in healthcare among lesbian, gay and bisexual (and other non-heterosexual) people is purported to be low, at least in some areas (e.g., cervical screening in lesbian and bisexual females, STI prevention in gay and bisexual males).^2–5 However, what is known more generally about rates of engagement in healthcare and interactions with health professionals among lesbian, gay and bisexual people, particularly in New Zealand, is limited.

Regular engagement with healthcare services is particularly important for marginalised populations, including lesbian, gay and bisexual people, given their elevated risk of both physical and mental health issues.⁶ For example, in both lesbian/bisexual females and gay/bisexual males, the incidence of cancers is higher than for cisgender heterosexual females and males.⁷ Lesbian and bisexual females are disproportionately affected by type 2 diabetes and ischaemic heart disease due to weight issues,^8,9 while stigma around gay male sexuality places gay and bisexual males at greater risk of sexual health issues.¹⁰ Due to systemic marginalisation, lesbian, gay and bisexual people experience higher rates of mental health issues (e.g., depression, anxiety) than do heterosexuals.¹¹ Systemic marginalisation based on gender and/or race especially affects the mental wellbeing of lesbian, gay and bisexual people who are also from ethnic minorities and/or who are gender diverse (transgender or non-binary).^12,13

Despite considerable social change, research with lesbian, gay and bisexual people consistently report problematic interactions with healthcare professionals. A recent systematic review of studies identified four main factors adversely impacting the healthcare experiences in this population.¹⁴ These factors include healthcare professionals’ apparent lack of knowledge of the specific health needs of this population; discomfort during interactions with lesbian, gay and bisexual patients; and the enactment of heteronormative assumptions, issues that result in culturally inappropriate care. For these reasons, healthcare professionals may fail to identify conditions specifically affecting lesbian, gay and bisexual people^15,16 and/or overlook the ways in which their health needs differ from those of heterosexuals.^17,18 While overt prejudice is rare in healthcare settings today, the persistence of heteronormativity is a key factor in many lesbian, gay and bisexual people being reticent to engage with—or, in some cases, actively avoiding—healthcare services.^19–21

While there are some studies that explore engagement with professionals and processes in primary healthcare in Aotearoa among lesbian, gay and bisexual people,^2,22 these have not utilised national level data. Using data collected from the New Zealand Health Survey (NZHS), the purpose of this study was to compare a) the engagement with general practitioner (GP) and nursing services, and b) patient experiences of GP services of lesbian, gay and bisexual people and heterosexual people. Rather than focussing on general experiences of healthcare—as has been the case in previous studies—this article explores experiences as defined in the NZHS; namely, levels of trust, explanations of health conditions and treatments and patient involvement in decision-making. Experiences of interactions were only explored for GPs, as the NZHS does not currently ask for this information in relation to nurses.

Method

Study design

The data analysed in this paper were extracted from the NZHS. The NZHS is a nationally representative survey that employs a multi-stage sampling method and conducts face-to-face interviews with respondents aged 15 years and older residing in Aotearoa New Zealand. We used data from the NZHS spanning four waves from the years 2017/2018 to 2020/2021, and combined all four waves of the survey, resulting in a total sample size of 46,849. The study reported here uses quantitative data to undertake a comparative analysis of lesbian/bisexual females, gay/bisexual males and heterosexual males and females. For the purposes of this article, only those identifying with the gender categories “male” or “female” were included in the analysis. The terms “male” and “female” are used throughout this article rather than “women” and “men”, as this is the language used in the NZHS to ask about gender. There is currently no provision in the survey to identify as “non-binary” or with a gender other than male or female.

Sample

Prior to analysis some data cleaning was necessary to ensure the sample only included people who had responded to all relevant questions. We excluded respondents who did not provide data for the main variables of interest: utilisation of GPs (n=70), utilisation of nurse service (n=163) and sexual identity (2,501). In total, we excluded 2,689 respondents who did not report those main outcomes (i.e., GP and nurse utilisation) and/or independent variables (i.e., sexual identity). Additionally, respondents who did not respond to variables used for adjustments in the model (e.g., income, education) were excluded (n=1,537). The final sample size for the analysis of GP and nurse service utilisation was 42,623.

Additionally, we conducted an analysis of patient experiences with GP services. In the NZHS, the assessment was limited to individuals who had visited a GP within the past 3 months, thereby focussing the analysis on this specific group (n=21,363). Respondents who did not provide answers to questions regarding their patient experience of GP services were excluded from the analysis (n=2,182). The total sample size for the analysis of experiences in GP services was 19,181.

Independent variables and dependent variables

Sexual identity

Sexual identity was self-defined and derived from responses to the survey question “Which of the following options best describes how you think of yourself?” The response options included heterosexual or straight, gay or lesbian, bisexual, other (not further defined), don’t know and choose not to answer. To enable a comparative analysis, we grouped individuals into the sexual identity categories “heterosexual females”, “lesbian”, “bisexual females”, “heterosexual males”, “gay” and “bisexual males”. Those who responded to either question with “don’t know” or “choose not to answer” or “other” were excluded from the analysis.

GP utilisation and experiences

We analysed the main outcome variables that were consistent across all four waves of the survey, which included the utilisation of GP services and patient experiences in GP services.

For GP utilisation, respondents were asked, “In the past 12 months, have you seen a GP or been visited by a GP about your own health? By health, I mean your mental health and emotional health as well as physical health.” Response options included yes, no, don’t know and refused.

Concerning patient experiences in GP services, respondents were queried about their experience if their last visit to a GP was within the last 3 months. There were three primary outcome variables related to patient experience in GP services: levels of trust, explanations of health conditions and treatments, and patient involvement in decision-making. Levels of trust in the GP were determined by asking respondents, “Did you have confidence and trust in the GP you saw or talked to?” Response options included yes, to some extent and no, not at all. Experiences regarding medical doctors explaining health conditions and treatments were assessed with the question, “Still thinking about your last visit or talk with a GP, how good was the doctor at explaining your health conditions and treatments in a way that you could understand?” Response options included very good, good, neither good nor bad, poor and very poor. The involvement of patients in decision-making about healthcare and treatment was evaluated by asking, “How good was the doctor at involving you in decisions about your care, such as discussing different treatment options?” The response options were the same as those mentioned for the second patient experience explained earlier.

Nurse utilisation

The nurse utilisation was measured using following question, “In the past 12 months, have you seen or talked to a nurse at a GP clinic or medical centre, about your own health?” Response options included yes, no, don’t know and refused. There was no question related to patient experience regarding nurse utilisation.

Analyses

Statistical analyses were undertaken using Stata version 11. Analyses comprised both univariate and multivariate calculations. Descriptive statistics were used to identify the percentage utilisation for both GPs and nurses. A Chi-squared test was utilised for a univariate analysis to examine differences in the percentage of GP utilisation and nurse utilisation across different sexual identities. For investigating the association between GP utilisation and nurse utilisation and sexual identity, we employed multiple logistic regression. The model was adjusted for potential confounding variables (e.g., demographic variables, socio-economic indicators, etc.). For examining the association between patient experience and sexual identity, we utilised ordered logistic regression. While acknowledging potential violations of assumptions of proportional odds assumptions, the large sample size supports the practicality of using ordered logistic regression.²³ Additionally, to account for errors stemming from the study design, we applied weight analysis in all our analyses. We analysed based on complete case analysis because missing data is less than 10%; therefore, it might not have a large effect on the findings.^24,25

We also adjusted for key confounding factors that were consistently assessed in all survey waves. These factors encompassed demographic variables, socio-economic indicators, the presence of health insurance, health condition and the survey year. Demographic variables consisted of age groups and ethnicity. Socio-economic status was evaluated based on household incomes over the past 12 months, highest educational attainment (categorised as no formal education, secondary, undergraduate, postgraduate and other) and employment status (categorised as not paid and looking for a job, not paid due to retirement and other groups). Health conditions took into account the presence of chronic diseases.

We also adjusted for gender and sexual identity in the model; however, the model was simplified due to the overlap in categories between gender and sexual identity with the primary sexual identity variable. As a result, we excluded both gender and sexual identity from the model and instead included the main sexual identity variable, which integrates both dimensions.

Results

Association between sexual identity and GP utilisation and nurse utilisation

The overall percentage of GP utilisation among the samples was 80%, while nurse utilisation was 54%. Upon exploring the percentage of GP utilisation across different sexual identities, we observed that bisexual females had the highest percentage (84%), followed by heterosexual females (83%), and the lowest percentage was found among bisexual males (70%). The distribution was similar for nurse utilisation. Statistically significant differences were observed in the percentage of GP and nurse utilisation across sexual identity groups (see Table 1).

View Table 1–3.

After adjusting for potential confounders, lesbian and bisexual females were more likely to utilise GP services compared to heterosexual females. For example, bisexual females were 1.81 times more likely to utilise GPs than heterosexual females (OR:1.81; 95% CI:1.79, 1.83; p-value <0.05). However, heterosexual males and gay/bisexual males were less likely to utilise GP services. For example, bisexual males were approximately 43% less likely to utilise GP services compared to heterosexual females (OR:0.57; 95% CI:0.56, 0.58; p-value <0.05). This is similar to nurse utilisation, except for gay males. Gay males were 1.03 times more likely to utilise nurses than heterosexual females (OR:1.03; 95% CI:1.01, 1.04; p-value <0.05) (see Table 2).

Association between sexual identity and patient experience in GP services

To identify associations, a total of 19,181 samples were included in the analysis, comprising 11,278 heterosexual females, 118 lesbian females, 296 bisexual females, 7,311 heterosexual males, 116 gay males and 62 bisexual males.

Level of trust

After adjusting for potential confounders, we observed that bisexual females as well as gay males were more likely to report poorer scores of trust in GPs compared to heterosexual females. For example, bisexual females were 1.32 times more likely to report poorer scores of trust in GPs compared to heterosexual females (OR:1.32; 95% CI:1.30, 1.34; p-value <0.05). However, we found that heterosexual males and lesbian females were less likely to report poorer scores of trust in GP services compared to heterosexual females. For example, lesbian females were 7% less likely to report poorer scores of trust compared to heterosexual females (OR:0.93; 95% CI:0.91, 0.96; p-value <0.05) (see Table 3).

Patient experiences in explanation of doctors on health conditions and treatment

After adjusting for potential confounders, we observed that all sexual identities were more likely to experience poorer scores of explanations from doctors regarding health conditions and treatment compared to heterosexual females. For instance, bisexual females were 1.50 times more likely to experience poorer scores of explanations compared to heterosexual females (OR: 1.50; 95% CI: 1.49–1.52; p-value <0.05). However, the association for gay males was not significant (see Table 3).

Level of involvement in decision making on healthcare

After adjusting for potential confounders, we observed that all sexual identities, except bisexual males, were more likely to experience lower levels of involvement in decision making on health compared to the heterosexual females. For instance, lesbian females were 1.28 times more likely to report lower levels of involvement in decision making on health (OR: 1.28; 95% CI: 1.25–1.30; p-value <0.05). However, bisexual males were 13% less likely to report lower levels of involvement in decision making on health (OR: 0.87; 95% CI: 0.85–0.89; p-value <0.05) (see Table 3).

Discussion

The purpose of this study was to explore levels of engagement and experiences of interactions with GPs (and nurses) among lesbian, gay and bisexual people compared to heterosexual people. In relation to the utilisation of GPs and nurses, lesbian and bisexual females were more likely to use GP or nurse services than heterosexual females. However, both heterosexual and gay/bisexual males were less likely to utilise GPs and nurse services compared to heterosexual females. Gay males and lesbian females were more likely to report poorer experiences in the explanation of health conditions and treatments, and more likely to report lower levels of involvement in decision making about their healthcare. Bisexual males and females were more likely to experience lower levels of trust in GPs compared to heterosexual females.

It is surprising that lesbian/bisexual females were more likely to use GPs and nurses than were heterosexual females given lower rates of engagement in cervical screening in this population.² However, this might suggest that lower engagement in healthcare among lesbian/bisexual females is restricted to gynaecological services and that this population is less reticent about engaging with health professionals over other concerns. It is also possible that, due to health disparities, lesbian and bisexual females experience serious illness at higher rates, necessitating more frequent engagement with health professionals. Additionally, studies on the health of lesbian and bisexual females typically include people who identify as transgender or gender diverse, whereas this study does not. Given that transgender and gender diverse people are reported to experience substantive challenges in accessing healthcare and therefore may actively avoid engagement, these findings may reflect the nature of the study sample.

In relation to men, the results of this study are consistent with a long-standing concern about men being much less likely to access healthcare than women. Men’s engagement with healthcare services is impacted by factors such as structural barriers (e.g., general practices appearing unwelcoming), internal barriers (e.g., masculinity and stoicism) and a lack of understanding of the role of healthcare providers (e.g., overlooking role of prevention).²⁶ Many gay and bisexual men also very carefully manage their healthcare encounters,¹⁷ including not disclosing their sexuality to their healthcare providers.^27,28 Such disclosures could be due to past or expected experiences of discomfort, negative attitudes and judgement and heteronormative assumptions demonstrated by healthcare professionals.¹⁴ Additionally, some gay and bisexual men dismiss links between sexuality and health²⁹ and only feel it is relevant for doctors to know about sexuality if a consultation is about sexual behaviours.^17,22

Consistent with previous research,^{14­­­­­–18} the findings of this study indicate that lesbian, gay and bisexual people have a poorer experience of GP services than do their heterosexual counterparts. This was evident across all three indicators asked about in the NZHS (trust of GPs, the explanation of health conditions and treatments and in decision making about healthcare options). While these indicators differ from those asked about in other studies, they may be attributable to factors such as GP discomfort during interactions with lesbian, gay and bisexual patients and knowledge gaps that may make conveying information in culturally appropriate ways more challenging. Without recognition of the ways in which the health needs of lesbian, gay and bisexual people differ from those of heterosexuals, it is difficult for GPs to provide appropriate multi-dimensional healthcare. Too often the onus is put on patients to disclose their sexuality and/or educate healthcare professionals about the ways in which their healthcare needs differ from those of heterosexuals.

Policy and practice implications

The findings of this study point to the need for GPs and nurses to be better educated about the specific health needs of lesbian, gay and bisexual people, in tandem with how to exercise culturally competent practice in respect of this population. To do this, health professionals need to work in ways that create trust, leading to better provider–patient relationships. Rather than treating lesbian, gay and bisexual people as “special cases”, this could be best accomplished through engagement that approaches all patients as if they may be sexually diverse rather than assuming heterosexuality. To upskill health practitioners in this, it is necessary to make LGBTQIA+ affirmative awareness training part of the in-service training of GPs and nurses.

Strengths and limitations

This study is based on national representative samples and pooled data from four waves of the NZHS. However, despite this study using the most robust data available for the health of lesbian, gay and bisexual populations in New Zealand, a potential limitation relates to the collection of information about the sexuality of respondents. Lesbian, gay and bisexual people have been found to be under-represented in population surveys,³⁰ and it is not possible to know how comfortable respondents in the survey were in identifying they were gay, lesbian or bisexual, or how accurate the responses were. This concern is pertinent given there is current concern among officials about people’s willingness to answer questions about sexuality in government surveys.¹⁷ In addition, the data collection processes assume a binary conceptualisation of gender (i.e., male or female), with no recognition of other ways of accounting for gender identities/expressions and the subsequent implications for sexual identities/labels.¹⁸ Therefore, it is possible some people with a non-male or non-female gender identity may not have taken part or have been forced to use a label they were not comfortable with. There is also the possibility of a social-desirability bias, in that participants may have been more inclined to provide more desirable responses to questions about patient experience.

Our analysis was based on secondary data, which resulted in small sample sizes for certain groups, such as bisexual males. Consequently, interpretations regarding these groups should be made with caution. Moreover, the validity of the ordinal scales, specifically levels of trust, explanations of health conditions and treatments and patient involvement in decision making, was not assessed in this study, due to reliance on secondary data. The utilisation of GPs and nurses may be affected by COVID-19; however, in our study we did not consider this issue.