Development of a culture of evidence based medicine depends on a body of research that draws from both qualitative and quantitative approaches.1 Recent BMJ articles have usefully questioned a stark polarity between qualitative and quantitative research and helped to demystify qualitative approaches.2 3 4 There has been little mention of ethnography, however, and little argument for its use in health research.I have examined some of these omissions, giving a broad indication of the nature of ethnography and arguing for its greater use within health care. I have given examples of ethnographic studies to suggest some of the issues that ethnography can help to explore, together with a brief outline of limitations of the approach.

We investigated the contributions of low energy expenditure and high energy intake to excessive weight gain in infants born to overweight mothers. The subjects were infants of 6 lean and 12 overweight mothers, recruited soon after birth. Total energy expenditure and metabolizable energy intake were measured with a new doubly labeled water method over a period of seven days when the infants were 3 months of age, and the postprandial metabolic rate was measured by indirect calorimetry when the infants were 0.1 and 3 months of age. The results were related to weight gain in the first year of life. No significant difference was observed between infants who became overweight by the age of one year (50 percent of infants born to overweight mothers) and those who did not, with respect to weight, length, skinfold thicknesses, metabolic rate at 0.1 and 3 months of age, and metabolizable energy intake at 3 months. However, total energy expenditure at three months of age was 20.7 percent lower in the infants who became overweight than in the other infants (means +/- SE, 256 +/- 27 and 323 +/- 12 kJ per kilogram of body weight per day; P less than 0.05). This difference could account for the mean difference in weight gain. These data suggest that reduced energy expenditure, particularly on physical activity, was an important factor in the rapid weight gain during the first year of life in infants born to overweight mothers.

The objective is to explore how clinical decisions are made in a cancer multidisciplinary team meeting (MDM). The study design is qualitative based on participant observation, in depth interviews, and questionnaires. The research setting was weekly cancer MDM which provides a forum for clinical debate for practitioners in the field of women's health, working within one Cancer Network in England. The participants were 53 practitioners attending a weekly MDM over a 4-month period. Analysis of nonparticipant observation data and practitioner interview narratives identified key influences on the work of the MDM, and in particular decision-making. The research identified three major influences on the conduct of the MDM. First, MDM discussions are dominated by those with surgical, medical, or diagnostic expertise with limited contributions from those with a nursing, palliative, or psychosocial background. Second, decision-making is shaped by an overriding need to comply with policy initiatives concerning the organization of diagnosis and treatment. The third influence is whether the patient is known or unknown to some degree by members of the MDM. Where there is preexisting knowledge of the patient, the discussion and decision is inclusive of a wider range of disciplines. Team working in these circumstances is an acknowledged source of satisfaction and motivation. Where the patient is not known, discussion concerns only the physical details necessary to make a diagnosis and contributions from the wider team (including those with knowledge of psychosocial care) are rare. Practitioners' sphere of expertise, Department of Health policy, and familiarity of the team with the patient are key factors in shaping decision-making in MDMs.

This article argues that participant observation is more than mere method and in need of greater theoretical attention. This is particularly true for its more participative forms, which are inseparable from assumptions about the role of the body in the generation of knowledge. Drawing on fieldwork experience, parallels are noted between participative observation and the clinical practice of nursing, for example, their reliance on physical involvement, their claims to experiential knowledge, and the associated theoretical assumptions they share, such as a reciprocity of perspective between subject and object. Such assumptions need to be examined if the knowledge learnt through participation is to carry-weight.

Objective: To estimate more accurately the age specific prevalence of Trichomonas vaginalis, Chlamydia trachomatis, Neisseria gonorrhoeae, and human papillomavirus infection (HPV) in indigenous women living in urban, rural, and remote areas of the "Top End" of the Northern Territory (NT). Design: Analysis of data obtained from two community based studies using self administered tampon specimens tested by polymerase chain reaction for sexually transmitted disease (STD). Data pertaining to the notifiable STDs (N gonorrhoeae and C trachomatis) were obtained from the NT health department. Patients: 1090 indigenous women (age range 12-73 years) were enrolled when they attended local community health centres, family planning clinics, and STD clinics. The majority attended clinics in their home community in the course of "well women's checks" which encourage women to undergo screening for a variety of general medical conditions. Results: The overall prevalence of T vaginalis, C trachomatis, N gonorrhoeae, and HPV was 0.25 (95% CI: 0.22-0.28), 0.11 (0.09-0.13), 0.17 (0.15-0.19), and 0.42 (0.37-0.48) respectively. Of the women found to be infected (excluding HPV), 25.5% had two or more of the above organisms detected. There was a statistically significant increase in the age specific prevalence of T vaginalis but a significant decrease with age for C trachomatis and HPV infection. There was no statistically significant change for N gonorrhoeae with age. Conclusions: STDs are hyperendemic in this population of indigenous women and the notification data significantly underestimate their prevalence. Distinct patterns of age specific prevalence were demonstrated, highlighting the need to tailor control strategies to specific epidemiological features. (Sex Transm Inf 1999;75:431-434)