Participants and setting

Two London (the Camden and Islington and the Enfield and Haringey) local research ethical committees approved the study. We recruited people attending their primary care physicians. 95% of the people in the UK are registered at general practices that serve as the first point of contact with health services. We approached a group of research general practices in a defined area of north London. None of the practices approached served only a LGB population nor did they advertise or offer special health care to LGB people. We asked consecutive people aged 18 to 75 attending these general practices to participate in the study and in each practice recruited attendees over a 4–8 week period. Each person approached to participate in the study was given a detailed information sheet on the study procedures. Those consenting to take part were asked to complete a questionnaire in a private setting in the general practice as they waited to see the doctor [16].

Definition of sexual orientation and physical, psychological and sexual problems

Analysis

We examined the data using descriptive statistics. Differences in the demographic, sexual activity, and health outcome data between gay/lesbian, bisexual and heterosexual classified people were examined using the Chi squared statistic for dichotomous and analysis of variance for continuous data. We then explored the influence of age, civil status, ethnicity and current employment on sexual orientation and the primary outcomes of interest. These were physical quality of life (as measured on SF-12) and psychological distress (on the GHQ-12), sexual function, number of sexual partners in the previous month, childhood sexual experiences, CAGE scores, smoking status and consultation rates. Only civil status (married or living with a partner versus the remainder) and ethnicity (white versus non white) were found to be associated both with sexuality and the health outcomes at a significance level of 10% or less. Thus, we adjusted the analysis of each health factor against our classification of sexual orientation for each of these confounders by fitting logistic models and comparing their fit using the likelihood ratio. We report the p values to indicate whether the adjustments made a significant difference to the model. People in the heterosexual group were used as the index population on the grounds of their overwhelming majority. Each multivariate analysis was conducted separately for men and women. We analysed the data using SPSS version 10 and Stata version 7.

Response rates and sexuality

Associations between health, sex and sexual orientation

Independent factors associated with health and well-being

After adjustment for civil and ethnic status, men classified as bisexual were more likely than those classified as heterosexual to be current smokers (OR 2.48), to report having had more than one sexual partner in the last four weeks (OR 5.0) and to be dissatisfied with their sex lives (OR 2.74) (table 3). Men classified as gay were more likely than those classified as heterosexual (OR 2.52) to score above the threshold of the general health questionnaire indicating current psychological distress and to report sexual experiences in childhood (OR 4.86) (table 3). Women classified as bisexual (OR 2.53) and lesbian (OR 3.13) were more likely to be current smokers than those classified as heterosexual (table 4). Women classified as bisexual were significantly more likely than those classified as heterosexual (OR 2.73) to record a positive CAGE score indicating possible alcohol misuse, to have had more than one sexual partner in the last four weeks (OR 2.85) and to report sexual experiences in childhood (OR 3.27) (table 4).

Sexual experiences in childhood and current pyshcolgical distress

Sexual experiences in childhood and high GHQ-12 scores appeared to occur together in gay and bisexual classified men (table 3) and so we explored their association further. To do this we compared heterosexual classified participants with gay and bisexual classified participants combined. This revealed associations only for gay and bisexual men and heterosexual women. Thirty-six per cent (18/50) of heterosexual classified men who reported childhood sexual experiences had GHQ-12 scores over 2 compared to 34% (107/315) of such men who did not report childhood experiences. Similarly, 48% (15/31) of lesbian and bisexual classified women who reported childhood experiences had GHQ-12 scores over 2 compared to 44% (35/79) of such women who did not report childhood experiences. In contrast, 65% (15/23) of gay and bisexual classified men who reported childhood sexual experiences had high GHQ-12 scores compared to 45% (n = 9/20) who did not report childhood experiences (Chi2 = 6.16, P = 0.013). And similarly, 51% (57/112) of heterosexual classified women who reported childhood sexual experiences had high GHQ-12 scores compared to 38% (306/808) who did not report such experiences (Chi2 = 6.98, P = 0.008).

Main findings

In comparison to their heterosexuals counterparts: 1) women classified as lesbian were more likely to be smokers; 2) men classified as gay had higher levels of psychological symptoms and were more likely to report childhood sexual experiences; 3) women classified as bisexual were more likely to misuse alcohol, to be smokers, to report more than one sexual partner in the preceding four weeks and to report childhood sexual experiences; 4) men classified as bisexual were more likely to report more than one sexual partner in the preceding four weeks and to be dissatisfied with their sex lives. Reporting childhood sexual experiences was associated with adult psychological distress in gay and bisexual classified men and heterosexual classified women.

Strengths and limitations of the study

To our knowledge this is the first European study in which the mental and physical health of people of a range of sexual orientation, attending family practitioners has been compared. Consecutive recruitment meant that our samples of gay and bisexual people were comparatively small and led to relative uncertainty in our estimates of odds ratios. In keeping with the higher rate of attendance of women compared with men in UK general practice [32], in this study we recruited just over twice the numbers of women rather than men. We did not find any differences in the consultation rates of attendees classified as gay, lesbian, bisexual or heterosexual. Nevertheless, it is still possible that differences in help seeking behaviour between men and women and between people classified as gay, lesbian and bisexual would make it difficult to generalise our finding to either sex or people classified as gay, lesbian or bisexual as a whole. Recruiting general practice attendees also means that the prevalence of physical or psychological difficulties may have been higher than in the general population, given that participants would often have been seeking help for medical or social problems. Our results are hence limited to the people recruited to our study and may not represent that of the UK or Europe as a whole.

We used a definition of sexuality based on criteria developed for this study. There is little consensus on how to measure sexual orientation. Several issues should be considered. Firstly, sexual responsiveness to others of the same sex, like most human traits is believed to be continuously distributed in the population [17, 33]. Secondly, it may be incorrect to presume that such traits are stable within each person over time [34]. Thirdly, conflating any same-sex experiences with a categorization of the person as homosexual may present limitations when defining sexuality. Lastly, defining sexuality solely on the basis of sexual experience [17] may exclude people who fantasize about sex with others of the same sex but never have sexual contact [35]. Modern concepts of sexual orientation consider personal identification, sexual behaviour and sexual fantasy [36]. Few studies, however, utilise all these three definitions in arriving at a composite categorization of sexuality. One widely established definition is a person "with an orientation towards people of the same gender in sexual behaviour, affection, or attraction, and/or self-identity as gay/lesbian or bisexual" [37]. Using these types of definition, there is evidence that at least five per cent of people in western countries are gay or bisexual [14, 34–38]. In our study classification of sexuality was limited by the Kinsey classification for which no time frame was applied. Moreover the classification of gay, bisexual and heterosexual as applied to people recruited in our study has not been previously used in similar population and it is possible that a different classification using the Kinsey scale might have yielded different results. Nevertheless, based on the categorization of sexual orientation used in this study, prevalence estimates of gay, lesbian and bisexual people in our sample closely matched that reported from London in a larger UK wide population survey [14].

There were also limitations in the way some data were collected. We assessed sexual behaviour using standardised questionnaires that were not explicit in their definition of sexual intercourse [27–29]. Many people (gay or straight) regard sexual intercourse differently and do not always realise that it means penetrative (vaginal or anal) sex. Furthermore, for gay men receptive anal intercourse can be regarded as "sexual intercourse" more often than insertive anal intercourse. At first sight our figures for sexual intercourse in gay classified men may seem low and those for lesbian classified woman high. However, most of our evidence on the prevalence of anal intercourse in gay men arises from particular populations, such as younger age groups, men attending clubs and bars or those living in gay neighbourhoods. Studies using diary records show that reported rates of intercourse depend on whether not the man has a regular same-sex partner and if so whether the relationship is open or closed. Those in closed relationships and those without regular partners may have relatively low rates of penile-anal intercourse [39]. Given the wide age range in our sample and the fact that these men were not selected on any sexual risk grounds, it is not surprising that only 35% of gay men reported sexual intercourse in the preceding 4 weeks. Moreover, vaginal penetrative practices using sex toys and fingers or hands are commonly reported by lesbian and bisexual women [40, 41].

Mental health of gay men and lesbians

Our finding of higher rates of common psychological symptoms of depression and anxiety, in gay classified men accords with those of a recent study of the mental health of gay men and lesbians in England and Wales in which a large sample was recruited by snowball sampling [12]. There are a number of reasons why gay people may be more likely to report psychological difficulties, which include difficulties growing up in a world orientated to heterosexual norms and values and the negative influence of social stigma against homosexuality. In addition, the gay commercial world in which some men and women may participate to find partners and friends may make misuse of alcohol and cigarettes more likely [3, 12]. The former in particular can have adverse effects on mental well-being. Finally, our results add to evidence that sexual experiences in childhood in men classified as gay or bisexual may play a role in adult psychological adjustment [42, 43]. Why this is not the case in lesbian and bisexual classified women requires further exploration.

Mental health of bisexual people

Much less is known about how bisexual, as opposed to gay people, accept their sexual orientation or whether the possibility of having partners of either sex introduces difficulties. Although there is anecdotal evidence that bisexuality is regarded negatively by gay and lesbian political and social groups, it may present to people living in a predominantly heterosexual world as being of a more acceptable social status than that of a gay or lesbian identity. We found, however, little difference in psychological distress between people classified as bisexual and heterosexual, despite the finding that women classified as bisexual reported more sexual experiences in childhood than heterosexual classified women. This finding has also been reported in a population-based study in North American, where bisexual (and lesbian) women had higher lifetime rates of substance misuse and coerced sex than heterosexual women [44].

Physical and sexual health

There were no differences between people classified as gay and heterosexual in terms of physical functioning based on a quality of life measure. People classified as bisexual on the other hand were more likely to be smokers raising concern about smoking related diseases such as cancers, cardiovascular diseases and chronic obstructive airway diseases. There is now good data on smoking among gay, lesbian and bisexual people. Data from population based health surveys and random student population samples in North America have suggested high rates of smoking in LGB people [38]. Men and women classified as bisexual were also more likely than heterosexuals to report contact with two or more sexual partners in the preceding month, potentially increasing their likelihood of acquiring sexually transmitted infections. For women classified as bisexual this combination with alcohol misuse could also place them at a greater risk of sexually transmitted infections. There is good evidence that risky sexual behaviour is increasing among men and women but comparative data on sexual behaviour of different sexual orientations has to our knowledge not been previously reported [14].

Family practice

There is varying evidence on the use of family practice by lesbian, gay and bisexual people. We have previously reported that LGB people consult general practitioners for emotional reasons more often than heterosexuals [12]. However, other research in the UK suggests that LGB people may avoid primary medical care [49] and those that do attend may not reveal their sexual orientation [50]. Awareness of a patient's sexual orientation may alert the family practitioner to potential difficulties, particularly with regard to alcohol misuse and smoking and to some extent sexual behaviour. There is a clear need for awareness on the part of family practitioners and openness on the part of patients about sexual orientation. This would allow practice staff adequate opportunities to monitor the smoking status, alcohol use, mental health, current numbers of sexual contacts and childhood sexual experiences in people of range of sexual orientation attending family practice. People at risk could then be offered early interventions to minimise their chances of developing more advanced illnesses.