• sexual health services
• women's views

As the paper on chlamydia screening by Dixon-Woods et al in this issue (p 335) shows, qualitative research can tell you how patients are referred to sexual health services and why they come. These provide compelling reasons for integrating qualitative with more standard quantitative methods in STI research. The control of infection depends not only on effective prevention and treatment of pathogens but also on the interactions between pathogen, host, and environment. Qualitative methodologies are integral to our understanding of at least two points in this triangle. Interview, fieldwork, and a range of other methods allow us to describe what people say and what they do, and to investigate the often complex relationships between the two.

The paper highlights findings from interviews with patients that are relevant to clinical and public health planning. The authors show that patients are “ambassadors” of the services and thus underscore the importance of lay rather than professional networks of referral to services. They demonstrate the importance of being able to phone for results and thus the need for patients to retain an element of control over interactions in the clinic. They show too that consultations are prompted less by concerns about a specific infection and more by a range of symptoms or behavioural cues that could fit with a number of infections, a generic “STI” as it were.

The major objection to qualitative research has generally been one of scale. STI clinic staff and scholars from other traditions may already “know” the findings of studies such as the ones reported in this issue; they seem to be mere common sense. Doctors, for example, already know that the stigma attached to clinics deters potential patients and that the fragmentation of services makes optimal management impossible, for example, in providing contraception alongside STI treatment. On the other hand, qualitative studies often also cause surprise and disbelief. In such situations, findings may be rejected above all for their lack of power: “the numbers were too small.” This problem has been addressed by treating the qualitative as a pilot study leading to larger scale investigations. For example, the relevance of health promotion for a generic, and possibly asymptomatic, STI as suggested by Dixon-Woods et al would need testing in a larger population. Yet, larger studies are not always necessary. It should be possible to introduce telephoning for results and then assess the benefits for some or all patients without further research.

A second standard method of dealing with issues of scale is to integrate qualitative and quantitative research more thoroughly. Research on sexuality requires particular sensitivity and large scale surveys have drawn upon initial contextual studies showing what can sensibly be asked and analysed in terms of what are often called “knowledge, attitudes, beliefs, and practices” (KABP). They have also drawn on subsequent fieldwork and interview to help interpret data.¹ A good example of combined qualitative and quantitative research is found in partner notification for STI. Recent studies in the United States have reported that sexual contact tracing provides a low yield in contrast with social contact tracing, most likely because it is hard to identify and contact all sexual partners on the one hand while, on the other, these partners are found largely among social groups of peers.² Research combining qualitative and quantitative techniques inside clinics, through interviews of social/sexual contacts and fieldwork at places where “high risk” individuals congregate has been combined in order to better understand transmission dynamics and design interventions. In research on risks of gonorrhoea, we tested a biological marker for unique sexual networks, differentiating types of gonorrhoea,³ so as to see whether it would be possible to generalise the results of in-depth interviews and fieldwork through potentially routine clinic data.

In STI research, the qualitative has further connotations. Two of the more important include the settings that are studied and the definitions of health employed. The study of Dixon-Woods and colleagues took place in an STI clinic but many have been conducted outside the clinic and often with marginal and deprived groups, sometimes known as the “hard to reach” or “out of reach” because they do not use health services. Secondly, qualitative research tends to involve a broad social definition of health, compared with a medical view of dysfunction. This may redress current biases towards overly technological understandings and provide more appropriate bases for health promotion as patients may be concerned with one infection only in the context of others, with STI in relation to reproductive health, or with sexual health in relation to general wellbeing. In research on sex work, we have attempted to combine qualitative and quantitative elements over several years so as to build up relationships with a stigmatised social group inside and outside the clinic and develop services considered appropriate by participants, based on this more holistic model of health. In developing a “quantified anthropology” and a “social epidemiology,” we also intended to provide insight into the content and meaning of activities associated with STI alongside objective outcome measures. Accordingly, neither the qualitative nor the quantitative element stands alone. If, for example, 20% of a “high risk” category do not use condoms all the time, it is necessary to understand when condoms are not used, among whom, and why before it is possible to design a feasible intervention.⁴

It should not be forgotten that the scale of qualitative research is also its strength, enabling an understanding of interactions between host, pathogen, and environment over time and in the context of relevant variables such as relationships with sexual partners and service use. These advantages tend to be lost in studies of large numbers and so it is preferable to employ one of the analytic techniques available for generalising findings about social relationships concerning class, network, sex, and so forth to “scale up” findings of this kind. Further development of these markers will make it possible to achieve greater integration between complementary studies of the biological and social, the qualitative and the quantitative.