Publication bias in qualitative research: what becomes of qualitative research presented at conferences?

BACKGROUND: Less than half of studies presented at conferences remain unpublished two years later, and these studies differ systematically from those that are published. In particular, the unpublished studies are less likely to report statistically significant findings, and this introduces publication bias. This has been well documented for quantitative studies, but has never been explored in relation to qualitative research. METHODS: We reviewed the abstracts of qualitative research presented at the 1998 (n = 110) and 1999 (n = 114) British Sociological Association (BSA) Medical Sociology meetings, and attempted to locate those studies in databases or by contacting authors. We also appraised the quality of reporting in each abstract. RESULTS: We found an overall publication rate for these qualitative studies of 44.2%. This is nearly identical to the publication rate for quantitative research. The quality of reporting of study methods and findings in the abstract was positively related to the likelihood of publication. CONCLUSION: Qualitative research is as likely to remain unpublished as quantitative research. Moreover, non-publication appears to be related to the quality of reporting of methodological information in the original abstract, perhaps because this is a proxy for a study with clear objectives and clear findings. This suggests a mechanism by which "qualitative publication bias" might work: qualitative studies that do not show clear, or striking, or easily described findings may simply disappear from view. One implication of this is that, as with quantitative research, systematic reviews of qualitative studies may be biased if they rely only on published papers.

Squamous cell carcinoma of the bladder presenting as vulvitis and cliteromegaly.

INTRODUCTION: Clitoral metastases are exceptionally rare. We present a case of a squamous cell carcinoma of the bladder presenting with a clitoral metastasis. CASE REPORT: We report the case of an 84-year-old lady with frequency, dysuria and a clitoral mass, which was found to be a poorly differentiated carcinoma on fine needle aspiration cytology. Cystoscopy revealed a moderate to poorly differentiated squamous cell carcinoma of the bladder. CONCLUSION: This is the first reported case of a squamous cell carcinoma of the bladder with clitoral metastasis. Clitoral metastases are exceptionally rare, with only seven previous cases reported in the literature. Although the commonest cause of cliteromegaly is hormonal, a metastatic carcinoma should be considered as part of the differential diagnosis in the elderly female.

Targeting of bladder cancer with monoclonal antibody NCRC48--a possible approach for intravesical therapy.

OBJECTIVE: To determine the localization of the anti-MUC1 mucin monoclonal antibody (mAb) NCRC48 to bladder cancer following intravesical administration. PATIENTS AND METHODS: mAb NCRC48 (330-500 micrograms) radiolabelled with 111indium (11-17 MBq) was administered intravesically to 12 unselected patients with radiological evidence of bladder cancer. Tumour localization was assessed by gamma-camera imaging and by tissue biodistribution studies on biopsies obtained at cystoscopy at about 2 or 24 h after the procedure. After 24 h, whole blood radioactivity was measured and 3 weeks after the procedure the serum level of human anti-mouse antibodies was estimated using an ELISA method. RESULTS: Eleven patients had tumours confirmed at cystoscopy (grades 1-3, stages pTa-pT2). The mean uptake of NCRC48 by tumour and by normal urothelium (expressed as the percentage of the instilled dose/g x 10(3) +/- SD) at 2 h was 3.42 +/- 3.68 and 0.41 +/- 0.77 (P < 0.05). After 24 h, the values for tumour and normal urothelium were 1.17 +/- 1.18 and 0.17 +/- 0.11, respectively. Areas of increased activity on the scintigrams were consistent with the position of the tumours at cystoscopy. No radioactivity was detected in blood at 24 h and there was no evidence of a human anti-mouse antibody response. CONCLUSION: The MUC1 mucin may be a suitable antigen to study the potential of therapeutic strategies based on monoclonal antibody targeting of superficial bladder cancer and may allow the development of more effective agents in the treatment of this condition.

Advice over the telephone.

The authors of this paper anonymously telephoned A&E departments in the Thames area that provided 24-hour emergency cover to investigate the safety and validity of medical advice given by these departments over the telephone to the public.

Emergency oesophageal transection for uncontrolled variceal haemorrhage.

Continued haemorrhage from oesophageal varices despite adequate injection sclerotherapy and tamponade has a high mortality rate. Such patients are usually referred for surgery. Over a 10-year period, 30 patients (21 men and nine women of median age 52 (range 21-70) years) with acute variceal haemorrhage uncontrolled by initial treatment underwent early emergency oesophageal transection. Portal hypertension was caused by alcoholic cirrhosis in 22 patients; other forms of cirrhosis were present in seven and portal vein thrombosis in one. Hepatic function immediately before operation was Pugh grade A in two patients, B in six and C in 22. Deterioration between admission and transection from grade A to B occurred in one patient and from B to C in five. Oesophageal transection stopped variceal haemorrhage in 29 of the 30 patients. Rebleeding from gastric varices within 35 days of surgery occurred in five patients. Postoperative haemorrhage also occurred from perioesophageal vessels (two patients), a gastrotomy (one) and oesophageal ulceration (two). Hepatic failure developed in seven patients, renal failure in five and both hepatic and renal failure in four. Mortality at 30 days occurred in neither of the two patients with liver function of grade A, in one of six of grade B and in 18 of 22 of grade C. The overall 30-day mortality rate was thus 63 per cent. Mortality was related to the preoperative Pugh grade (hazard ratio 3.95 per grade; P = 0.013) and preoperative blood transfusion (hazard ratio 1.37 per unit; P = 0.035). Four of six patients with grade B liver function died within 3 months and 21 of 22 with grade C disease within 1 year. Oesophageal transection is effective at stopping variceal bleeding but does not modify the underlying disease. Caution is urged for patients with grade C hepatocellular impairment proceeding to acute oesophageal transection after initial sclerotherapy. Such patients may benefit more from treatment with somatostatin or an intrahepatic porta-systemic stent shunt while awaiting definitive therapy.

Primary pneumococcal peritonitis.

Primary pneumococcal peritonitis is an uncommon condition which in the preantibiotic era was associated with a high mortality rate. We present two cases and discuss the aetiology, clinical features and management of this condition.

PIP: Primary pneumococcal peritonitis is an uncommon condition 1st identified in 1885. It occurs when peritoneal inflammation is present in the absence of an intraabdominal source of infection. In the preantibiotic era, the condition accounted for 2% of childhood abdominal emergencies largely among girls aged 2-10 years. Mortality was 42-100%, with death sometimes occurring within 48 hours of the onset of symptoms. This condition now present in female adults, is associated with IUD use, and is comparatively common in India. Consideration should therefore be given to the existence of primary pneumococcal peritonitis when diagnosing and managing abdominal emergencies. The pneumococcus may enter the peritoneal cavity via the female genital tract, blood, or through transmural spread from the gastrointestinal tract. No evidence supports a relationship between type of IUD and/or length of time in place, and the onset of peritonitis. Given pneumococcus' commensal existence in the upper respiratory tract, urogenital sex may facilitate its entrance to the peritoneal cavity through the female genital tract. Abdominal pain, diarrhea, and vomiting generally present, while the patient may also be pyrexial and dehydrated. In diagnosing this condition, the practitioner may confuse it with acute appendicitis, pelvic inflammatory disease, or gastroenteritis if in the early stages of peritonitis. Diagnosis is often confirmed only thorough laparotomy, but abdominal paracentesis and/or abdominal ultrasound may also be employed as diagnostic aids. Laparotomy and a regime of antibiotics is the preferred treatment. 2 case studies are discussed.