A prospective investigation of oral contraceptive use and breast cancer mortality: findings from the Swedish women's lifestyle and health cohort.

BACKGROUND: The association between oral contraceptive (OC) use and long-term mortality remains uncertain and previous studies have reported conflicting findings. We aim to assess the long-term impact of OC use on all-cause and cancer-specific mortality. METHODS: Out of 49,259 participants, we analysed data on 2120 (4.3%) women diagnosed with first primary breast cancer between 1993 and 2012, in the Swedish Women's Lifestyle and Health Study. Kaplan-Meier plots were used to graph the hazard of mortality in association with oral contraceptives use, stage of disease and hormone receptors status at diagnosis. Cox proportional hazard model were used to estimate hazard ratios (HR) between OC use and all-cause mortality. The same association was studied for breast cancer-specific mortality by modelling the log cumulative mortality risk, adjusting for clinical stage at diagnosis, hormone receptor status, body mass index and smoking. RESULTS: Among 2120 women with breast cancer, 1268 (84%) reported ever use of OC and 254 died within 10 years of diagnosis. The risk of death for OC ever-users relative to never-users was: HR = 1.13 (95% CI: 0.66-1.94) for all-cause mortality and HR = 1.29 (95% CI: 0.53-3.18) for breast cancer-specific mortality. A high percentage of women (42.9%) were diagnosed at early stage disease (stage I). CONCLUSIONS: Among women with primary breast cancer, OC ever-users compared to never- users did not have a higher all-cause or breast cancer specific-mortality, after the adjustment of risk factors.

Primary school students' mental health in Uganda and its association with school violence, connectedness, and school characteristics: a cross-sectional study.

BACKGROUND: Few studies have explored risk factors for poor mental health in Ugandan primary schools. This study investigated whether individual- and contextual-level school-related factors including violence from school staff and other students, connectedness to school and peers, as well as school size and urban/rural location, were associated with mental health difficulties in Ugandan children. We also examined whether associations between violence exposure at school and mental health were mediated by connectedness as well as whether associations were different for boys and girls. METHODS: The analytic sample consisted of 3,565 students from 42 primary schools participating in the Good Schools Study. Data were collected through individual interviews conducted in June and July 2012. Mental health was measured using the Strengths and Difficulties Questionnaire. Multilevel logistic regression was applied to investigate factors associated with mental health difficulties. RESULTS: Experiences of violence from school staff and other students in the past week were strongly associated with mental health difficulties (OR = 1.58, 95 % CI 1.31 to 1.90 and 1.81, 1.47 to 2.23, respectively). Children with a low school connectedness had 1.43 times (1.11 to 1.83) the odds of mental health difficulties compared to those with a high school connectedness. The OR comparing children never feeling close to other students at their school with those always feeling close was 1.86 (1.18 to 2.93). The effect of violence on mental health was not mediated through the connectedness variables. School size was not related to mental health difficulties, but attending an urban school increased the odds of mental health difficulties after accounting for other factors. We did not find evidence that the effect of one or more of the exposures on the outcome differed between boys and girls. CONCLUSIONS: These findings suggest that violence in school and low connectedness to school and peers are independently associated with mental health difficulties and interventions should address both concurrently. Extra support may be needed for students in urban schools. TRIAL REGISTRATION: Clinicaltrials.gov NCT01678846 . Registered 24 August 2012.

Estimating excess hazard ratios and net survival when covariate data are missing: strategies for multiple imputation.

BACKGROUND: Net survival is the survival probability we would observe if the disease under study were the only cause of death. When estimated from routinely collected population-based cancer registry data, this indicator is a key metric for cancer control. Unfortunately, such data typically contain a non-negligible proportion of missing values on important prognostic factors (eg, tumor stage). METHODS: We carried out an empirical study to compare the performance of complete records analysis and several multiple imputation strategies when net survival is estimated via a flexible parametric proportional hazards model that includes stage, a partially observed categorical covariate. Starting from fully observed cancer registry data, we induced missingness on stage under three scenarios. For each of these scenarios, we simulated 100 incomplete datasets and evaluated the performance of the different strategies. RESULTS: Ordinal logistic models are not suitable for the imputation of tumor stage. Complete records analysis may lead to grossly misleading estimates of net survival, even when the missing data mechanism is conditionally independent of survival time given the covariates and the bias on the excess hazard ratios estimates is negligible. CONCLUSIONS: As key covariates are unlikely missing completely at random, studies estimating net survival should not use complete records. When the missingness can be inferred from available data, appropriate multiple imputation should be performed. In the context of flexible parametric proportional hazards models with a partially observed stage covariate, a multinomial logistic imputation model for stage should be used and should include the Nelson-Aalen cumulative hazard estimate and the event indicator.

Population-based cancer survival trends in England and Wales up to 2007: an assessment of the NHS cancer plan for England.

BACKGROUND: The National Health Service (NHS) cancer plan for England was published in 2000, with the aim of improving the survival of patients with cancer. By contrast, a formal cancer strategy was not implemented in Wales until late 2006. National data on cancer patient survival in England and Wales up to 2007 thus offer the opportunity for a first formal assessment of the cancer plan in England, by comparing survival trends in England with those in Wales before, during, and after the implementation of the plan. METHODS: We analysed population-based survival in 2.2 million adults diagnosed with one of 21 common cancers in England and Wales during 1996-2006 and followed up to Dec 31, 2007. We defined three calendar periods: 1996-2000 (before the cancer plan), 2001-03 (initialisation), and 2004-06 (implementation). We estimated year-on-year trends in 1-year relative survival for patients diagnosed during each period, and changes in those trends between successive periods in England and separately in Wales. Changes between successive periods in mean survival up to 5 years after diagnosis were analysed by country and by government office region of England. Life tables for single year of age, sex, calendar year, deprivation category, and government office region were used to control for background mortality in all analyses. FINDINGS: 1-year survival in England and Wales improved for most cancers in men and women diagnosed during 1996-2006 and followed until 2007, although not all trends were significant. Annual trends were generally higher in Wales than in England during 1996-2000 and 2001-03, but higher in England than in Wales during 2004-06. 1-year survival for patients diagnosed in 2006 was over 60% for 12 of 17 cancers in men and 13 of 18 cancers in women. Differences in 3-year survival trends between England and Wales were less marked than the differences in 1-year survival. North-South differences in survival trends for the four most common cancers were not striking, but the North West region and Wales showed the smallest improvements during 2001-03 and 2004-06. INTERPRETATION: The findings indicate slightly faster improvement in 1-year survival in England than in Wales during 2004-06, whereas the opposite was true during 2001-03. This reversal of survival trends in 2001-03 and 2004-06 between England and Wales is much less obvious for 3-year survival. These different patterns of survival suggest some beneficial effect of the NHS cancer plan for England, although the data do not so far provide a definitive assessment of the effectiveness of the plan.

Impact of severity of personality disorder on the outcome of depression.

The influence of severity of personality disorder on outcome of depression is unclear. Four hundred and ten patients with depression in 9 urban and rural communities in Finland, Ireland, Norway, Spain and the United Kingdom, were randomised to individual problem-solving treatment (n=121), group sessions on depression prevention (n=106) or treatment as usual (n=183). Depressive symptoms were recorded at baseline, 6 and 12 months. Personality assessment was performed using the Personality Assessment Schedule and analysed by severity (no personality disorder, personality difficulty, simple personality disorder, complex personality disorder). Complete personality assessments were performed on 301 individuals of whom 49.8% had no personality disorder; 19.3% had personality difficulties; 13.0% had simple personality disorder; and 17.9% had complex personality disorder. Severity of personality disorder was correlated with Beck Depression Inventory (BDI) scores at baseline (Spearman's r=0.21; p<0.001), 6 months (r=0.14; p=0.02) and 12 months (r=0.21; p=0.001). On multi-variable analysis, BDI at baseline (p<0.001) and type of treatment offered (individual therapy, group therapy, treatment as usual) (p=0.01) were significant independent predictors of BDI at 6 months. BDI at baseline was the sole significant independent predictor of BDI at 12 months (p<0.001). There was no interaction between personality disorder and treatment type for depression. While multi-variable analyses indicate that depressive symptoms at baseline are the strongest predictor of depressive symptoms at 6 and 12 months, the strong correlations between severity of personality disorder and depressive symptoms make it difficult to establish the independent effect of personality disorder on outcome of depression.

Effect of prepregnancy maternal BMI on adverse pregnancy and neonatal outcomes: results from a retrospective cohort study of a multiethnic population in Qatar.

BACKGROUND: Given the small number of studies on the topic, we aimed to identify the impact of prepregnancy maternal body mass index (BMI) on adverse pregnancy outcomes (POs) in a low-risk, multiethnic population, and to calculate related population attributable fractions (PAFs). METHODS: This retrospective cohort study included 1134 nulliparous women of 50 nationalities (classified into Arab and non-Arab ethnicity) in Qatar who had their first antenatal visit at a Primary Healthcare Corporation (PHCC) facility in June 2016-March 2017 and their PO at a Hamad Medical Corporation facility before 10 November 2017. We used multiple imputation to handle missing values and multivariate logistic regression to calculate adjusted ORs (aORs) for adverse POs in overweight and women with obesity. RESULTS: Overweight Arab women and women with obesity were at high risk for gestational diabetes mellitus (GDM) (aOR=2.38, 95% CI 1.51 to 3.84) and caesarean section (aOR=1.57, 95% CI 1.00 to 2.48). Non-Arab women with obesity were at high risk for pre-eclampsia (aOR=3.83, 95% CI 1.00 to 15.00). PAFs showed that 41.63% of pre-eclampsia, 17.36% of pregnancy-induced hypertension, 17.17% of large for gestational age, 15.89% of preterm deliveries, 14.75% of GDM and 13.99% of caesarean sections could be avoided if all mothers had normal prepregnancy BMI. There were no major differences in PAFs by ethnicity. CONCLUSION: Adverse POs were attributable to maternal obesity. This suggests that, in contrast to existing PHCC protocol, overweight and women with obesity in Qatar should be targeted earlier in their pregnancy; preferably prior to getting pregnant. We observed ethnic differences in the risk of adverse POs.

Oral Cancer Knowledge, Attitude, and Practice of Dentists in the State of Qatar.

OBJECTIVES: The aim of this study was to assess existing knowledge on oral cancer (OC), attitude toward OC examination, and clinical practice among dentists practicing in the governmental health sector in the State of Qatar, including the influence of personal characteristics. MATERIALS AND METHODS: All 271 dentists practicing in Primary Health Care Centers (PHCC) and the Hamad Medical Corporation (HMC) were invited to participate in this cross-sectional study. Of these, 177 completed a self-administered, 48-item questionnaire. Based on the responses, knowledge of the risk factors for and clinical presentation of OC was categorized as high, medium, or low, and then further separated into satisfactory (medium/high) and unsatisfactory (low). Information on attitudes toward OC examination and clinical practice was also taken from the questionnaire. RESULTS: The mean score for knowledge of the clinical presentation of OC was 7.59 (standard deviation [SD] = 2.40) out of 14. The mean score for knowledge of the risk factors for OC was 8.96 (SD = 2.31) out of 16. Dentists with ≤10 years of experience were more likely to have satisfactory knowledge of OC compared to dentists with >15 years of experience. Attending a continuous professional development (CPD) course on OC showed a trend with satisfactory clinical knowledge, although it was not statistically significant. CONCLUSION: This study identified gaps in dentists' knowledge of OC; dentists demonstrated unsatisfactory knowledge of the clinical presentation of and risk factors for OC. The findings highlighted the need for educational interventions on OC, which are essential to improving health care outcomes and delivery of care.

Optimal dilation time for combined small endoscopic sphincterotomy and balloon dilation for common bile duct stones: a multicentre, single-blinded, randomised controlled trial.

BACKGROUND: Endoscopic sphincterotomy is the established treatment for common bile duct stones. Balloon dilation offers an alternative. Prolonged dilation (300 s) with a 10 mm diameter balloon decreases the occurrence of pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP). We aimed to determine the optimal duration of dilation for combined endoscopic sphincterotomy and balloon dilation for the removal of common bile duct stones. METHODS: We did a multicentre, single-blinded, randomised controlled trial at 15 tertiary surgical centres in China. Eligible patients (≥18 years) with native papilla and common bile duct stones (≤1·5 cm in size and <2 cm in diameter) undergoing ERCP were randomly assigned (1:1:1:1:1) to receive balloon dilation for 0, 30, 60, 180, or 300 s after deep bile duct cannulation. Randomisation was done by an independent statistician using a computer-generated randomisation list with a block size of ten, stratified by centre. Patients and outcome assessors, but not endoscopists and investigators, were masked to treatment allocation. Balloon dilation was done with controlled radial expansion balloons according to common bile duct stone size. Stones were removed using stone retrieval balloons or baskets. The primary endpoint was overall frequency of post-ERCP pancreatitis. The primary efficacy analysis and safety analyses were done in the modified intention-to-treat population, which included all randomly assigned patients with successful cannulation, but excluded those who withdrew consent after randomisation. This study was registered with ClinicalTrials.gov, number NCT02510495, and is complete. FINDINGS: Between July 29, 2015, and Dec 1, 2017, 3721 consecutive patients with common bile duct stones were recruited, 1718 of whom were excluded. The remaining 2003 patients underwent a small (3-5 mm) endoscopic sphincterotomy. 83 patients withdrew consent after the ERCP procedure, thus 1920 patients were included in the modified intention-to-treat analysis (0 s [n=371], 30 s [n=384], 60 s [n=388], 180 s [n=390], and 300 s [n=387]). Overall, post-ERCP pancreatitis occurred in 199 (10%) of 1920 patients (44 [12%] patients in the 0 s group, 28 [7%] in the 30 s group, 32 [8%] in the 60 s group, 36 [9%] in the 180 s group, and 59 [15%] in the 300 s group). Prolonged dilation (300 s) significantly increased the occurrence of post-ERCP pancreatitis compared with shorter balloon dilation (p=0·002). The frequency of post-ERCP pancreatitis was significantly lower in the 30, 60, and 180 s groups than in the 300 s group (relative risk [RR] 0·48, 95% CI 0·31-0·73; p=0·0005 vs the 30 s group; 0·54, 0·36-0·81; p=0·003 vs the 60 s group; 0·61, 0·41-0·89; p=0·01 vs the 180 s group). The frequency of post-ERCP pancreatitis was significantly higher in the 0 s group than the 30 s group (RR 1·62, 1·04-2·56; p=0·03). No difference in stone extraction (all ≥90%) was observed between groups. Following ERCP, 90 (5%) of 1920 patients had acute cholangitis, 14 (<1%) had acute cholecystitis, and five (<1%) had gastrointestinal bleeding, with no significant differences between groups. One (<1%) patient had Stapfer II perforation, which resolved spontaneously with conservative treatment. INTERPRETATION: A balloon dilation time of 30 s for combined endoscopic sphincterotomy and balloon dilation reduced the frequency of post-ERCP pancreatitis and was determined to be the optimum dilation time for the removal of common bile duct stones. FUNDING: National Natural Science Foundation of China, Gansu Competitive Foundation Projects for Technology Development and Innovation.

Cognitive-behavioural therapy for health anxiety in a genitourinary medicine clinic: randomised controlled trial.

BACKGROUND: Little is known about the management of health anxiety and hypochondriasis in secondary care settings. AIMS: To determine whether cognitive-behavioural therapy (CBT) along with a supplementary manual was effective in reducing symptoms and health consultations in patients with high health anxiety in a genitourinary medicine clinic. METHOD: Patients with high health anxiety were randomly assigned to brief CBT and compared with a control group. RESULTS: Greater improvement was seen in Health Anxiety Inventory (HAI) scores (primary outcome) in patients treated with CBT (n=23) than in the control group (n=26) (P=0.001). Similar but less marked differences were found for secondary outcomes of generalised anxiety, depression and social function, and there were fewer health service consultations. The CBT intervention resulted in improvements in outcomes alongside higher costs, with an incremental cost of pound 33 per unit reduction in HAI score. CONCLUSIONS: Cognitive-behavioural therapy for health anxiety within a genitourinary medicine clinic is effective and suggests wider use of this intervention in medical settings.

A randomized controlled trial of cognitive-behavior therapy for persistent symptoms in schizophrenia: a five-year follow-up.

Meta-analyses of randomized controlled trials support the efficacy of cognitive behavioral therapy (CBT) in the treatment of symptoms of schizophrenia refractory to antipsychotic medication. This article addresses the issue of medium term durability. A five-year follow-up was undertaken of a sample of 90 subjects who participated in a randomized controlled trial of CBT and befriending (BF). Patients received routine care throughout the trial and the follow-up period. Intention to treat multivariate analysis was performed by an independent statistician following multiple imputation of missing data. Fifty-nine out of ninety patients were followed up at 5 years (CBT=31, BF=28). In comparison to BF and usual treatment, CBT showed evidence of a significantly greater and more durable effect on overall symptom severity (NNT=10.36, CI -10.21, 10.51) and level of negative symptoms (NNT=5.22, CI -5.06 -5.37). No difference was found between CBT and BF on either overall symptoms of schizophrenia or depression. The initial cost of an adjunctive course of CBT for individuals with medication refractory schizophrenia may be justified in light of symptomatic benefits that persist over the medium term.

Is age of menarche among school girls related to academic performance?

Background There is strong evidence that the mean age of menarche has declined over the last few decades in developed and developing countries. This is of a major concern because of its enormous public health implications. This study aimed to estimate the age of menarche in Kuwait and investigate the association between menarcheal age and academic performance among high school girls in Kuwait. Methods A cross-sectional study was conducted on randomly selected female high school students from private and public high schools in all governorates in Kuwait. Data on the age of menarche were collected by self-administered questionnaire from the students, while data on academic performance were extracted from the students' academic records. Results Of the 907 students we selected, 800 (88.2%) responded. The mean age of menarche was 12.33 [95% confidence interval (CI) 12.18-12.49] years. There was no evidence for significant association between age of menarche and students' academic performance before or after adjusting for potential confounders. Conclusion The calculated age of menarche among contemporary girls in Kuwait is similar to that of the girls in industrialized countries. Early menarcheal age is unlikely to lead to adverse behavior that may affect academic performance in our setting.

The Social Functioning Questionnaire: a rapid and robust measure of perceived functioning.

BACKGROUND: The Social Functioning Questionnaire (SFQ), an eight-item self-report scale (score range 0-24), was developed from the Social Functioning Schedule (SFS), a semi-structured interview which has been used primarily with non-psychotic patients and has good test-retest and inter-rater reliability as well as construct validity. The SFQ was developed following the need for a quick assessment of perceived social function. AIMS: To give further details of old and new data sets from studies involving over 4000 subjects assessed with the SFQ illustrating its epidemiological and clinical associations. METHOD: New data were analysed from a national epidemiological study, a comparison of key-worker and subject versions of the SFQ, and reanalysis of data from three earlier clinical studies, of psychiatric emergencies, general practice psychiatric patients and those with recurrent psychotic illnesses. These data were examined further to determine their range, their relationship to other clinical measures, and change over time in clinical trials. RESULTS: The population mean score in 4164 subjects was 4.6 and the data from all studies suggested that a score of 10 or more indicated poor social functioning. Those presenting as psychiatric emergencies had the poorest social function (mean 11.4) and psychiatric patients from general practice the best function (mean 7.7) of the clinical populations. The eight item scores had a normal distribution in psychiatric populations and a skewed one in a normal population; scores were relatively stable over the short (weeks) and long-term (months), and were high in the presence of acute mental health disturbance and personality disorder, giving support to the validity of the scale. The results from a UK sample of a randomly selected population specifically weighted for ethnic minorities showed similar social function across groups.

The impact of age at diagnosis on socioeconomic inequalities in adult cancer survival in England.

BACKGROUND: Understanding the age at which persistent socioeconomic inequalities in cancer survival become apparent may help motivate and support targeting of cancer site-specific interventions, and tailoring guidelines to patients at higher risk. PATIENTS AND METHODS: We analysed data on more than 40,000 patients diagnosed in England with one of three common cancers in men and women, breast, colon and lung, 2001-2005 with follow-up to the end of 2011. We estimated net survival for each of the five deprivation categories (affluent, 2, 3, 4, deprived), cancer site, sex and age group (15-44, 45-54, 55-64, and 65-74 and 75-99 years). RESULTS: The magnitude and pattern of the age specific socioeconomic inequalities in survival was different for breast, colon and lung. For breast cancer the deprivation gap in 1-year survival widened with increasing age at diagnosis, whereas the opposite was true for lung cancer, with colon cancer having an intermediate pattern. The 'deprivation gap' in 1-year breast cancer survival widened steadily from -0.8% for women diagnosed at 15-44 years to -4.8% for women diagnosed at 75-99 years, and was the widest for women diagnosed at 65-74 years for 5- and 10-year survival. For colon cancer in men, the gap was widest in patients diagnosed aged 55-64 for 1-, 5- and 10-year survival. For lung cancer, the 'deprivation gap' in survival in patients diagnoses aged 15-44 years was more than 10% for 1-year survival in men and for 1- and 5-year survival in women. CONCLUSION: Our findings suggest that reduction of socioeconomic inequalities in survival will require updating of current guidelines to ensure the availability of optimal treatment and appropriate management of lung cancer patients in all age groups and older patients in deprived groups with breast or colon cancer.

Inequalities in non-small cell lung cancer treatment and mortality.

BACKGROUND: Non-small cell lung cancer (NSCLC) comprises approximately 85% of all lung cancer cases, and surgery is the preferred treatment for patients. The National Health Service established Primary Care Trusts (PCTs) in 2002 to manage local health needs. We investigate whether PCTs with a lower uptake of surgical treatment are those with above-average mortality 1 year after diagnosis. The applied methods can be used to monitor the performance of any administrative bodies responsible for the management of patients with cancer. METHODS: All adults diagnosed with NSCLC lung cancer during 1998-2006 in England were identified. We fitted mixed effect logistic models to predict surgical treatment within 6 months after diagnosis, and mortality within 1 year of diagnosis. RESULTS: Around 10% of the NCSLC patients received curative surgery. Older deprived patients and those who did not receive surgery had much higher odds of death 1 year after being diagnosed with cancer. In total, 69% of the PCTs were below the lower control limit of surgery and have predicted random intercepts above the mean value of zero of the random effect for mortality, whereas 40% were above the upper control limit of mortality within 1 year. CONCLUSIONS: Our main results suggest the presence of clear geographical variation in the use of surgical treatment of NSCLC and mortality. Mixed-effects models combined with the funnel plot approach were useful for assessing the performance of PCTs that were above average in mortality and below average in surgery.

Control of data quality for population-based cancer survival analysis.

BACKGROUND: Population-based cancer survival is an important measure of the overall effectiveness of cancer care in a population. Population-based cancer registries collect data that enable the estimation of cancer survival. To ensure accurate, consistent and comparable survival estimates, strict control of data quality is required before the survival analyses are carried out. In this paper, we present a basis for data quality control for cancer survival. METHODS: We propose three distinct phases for the quality control. Firstly, each individual variable within a given record is examined to identify departures from the study protocol; secondly, each record is checked and excluded if it is ineligible or logically incoherent for analysis; lastly, the distributions of key characteristics in the whole dataset are examined for their plausibility. RESULTS: Data for patients diagnosed with bladder cancer in England between 1991 and 2010 are used as an example to aid the interpretation of the differences in data quality. The effect of different aspects of data quality on survival estimates is discussed. CONCLUSIONS: We recommend that the results of data quality procedures should be reported together with the findings from survival analysis, to facilitate their interpretation.

Socio-economic inequalities in testicular cancer survival within two clinical studies.

BACKGROUND: Testicular cancer is the most common cancer in men under 35 years of age, and has the highest survival for adult male malignancies. Despite the fact that survival is very high, there is evidence that survival differs between socio-economic groups. METHODS: We analysed survival patterns for 1606 testicular cancer patients diagnosed during 1984-2001 and recruited to one of two clinical studies. The first was a surveillance study to determine relapse-free survival after orchidectomy in 865 patients with stage I nonseminomatous germ-cell testicular cancer diagnosed during 1984-1991 (TE04). The second study was a trial in which 1174 men with stage I seminomatous germ-cell tumours were randomised to receive radiotherapy or one injection of carboplatin between 1996 and 2001 (TE19). The number of men available for analysis from these two studies was 578 and 1028, respectively. We followed these patients up for their vital status, and assigned them an ecological measure of deprivation. Crude and relative survival were estimated at 5 and 10 years by socio-economic deprivation. RESULTS: No significant socio-economic gradient was seen: 1.3% (95% CI -0.3% to 3.1%) at 5 years and 2.1% (95% CI -0.5% to 4.7%) at 10 years. CONCLUSION: We conclude that, given equal treatment at a given stage of disease, survival from testicular cancer does not depend on socio-economic status. This suggests that the socio-economic gradient in testicular cancer survival in the general population is more likely to be attributable to health care system factors than to personal or socio-economic factors in the men themselves.

Modelling relative survival in the presence of incomplete data: a tutorial.

BACKGROUND: Missing data frequently create problems in the analysis of population-based data sets, such as those collected by cancer registries. Restriction of analysis to records with complete data may yield inferences that are substantially different from those that would have been obtained had no data been missing. 'Naive' methods for handling missing data, such as restriction of the analysis to complete records or creation of a 'missing' category, have drawbacks that can invalidate the conclusions from the analysis. We offer a tutorial on modern methods for handling missing data in relative survival analysis. METHODS: We estimated relative survival for 29 563 colorectal cancer patients who were diagnosed between 1997 and 2004 and registered in the North West Cancer Intelligence Service. The method of multiple imputation (MI) was applied to account for the common example of incomplete stage at diagnosis, under the missing at random (MAR) assumption. Multivariable regression with a generalized linear model and Poisson error structure was then used to estimate the excess hazard of death of the colorectal cancer patients, over and above the background mortality, adjusting for significant predictors of mortality. RESULTS: Incomplete information on stage, morphology and grade meant that only 55% of the data could be included in the 'complete-case' analysis. All cases could be included after indicator method (IM) or MI method. Handling missing data by MI produced a significantly lower estimate of the excess mortality for stage, morphology and grade, with the largest reductions occurring for late-stage and high-grade tumours, when compared with the results of complete-case analysis. CONCLUSION: In complete-case analysis, almost 50% of the information could not be included, and with the IM, all records with missing values for stage were combined into a single 'missing' category. We show that MI methods greatly improved the results by exploiting all the information in the incomplete records. This method also helped to ensure efficient inferences about survival were made from the multivariate regression analyses.

Music therapy for in-patients with schizophrenia: exploratory randomised controlled trial.

BACKGROUND: Music therapy may provide a means of improving mental health among people with schizophrenia, but its effects in acute psychoses have not been explored. AIMS: To examine the feasibility of a randomised trial of music therapy for inpatients with schizophrenia, and explore its effects on mental health. METHOD: Up to 12 weeks of individual music therapy plus standard care were compared with standard care alone. Masked assessments of mental health, global functioning and satisfaction with care were conducted at 3 months. RESULTS: Of 115 eligible patients 81 (70%) were randomised. Two-thirds of those randomised to music therapy attended at least four sessions (median attendance, eight sessions). Multivariate analysis demonstrated a trend towards improved symptom scores among those randomised to music therapy, especially in general symptoms of schizophrenia. CONCLUSIONS: A randomised trial of music therapy for in-patients with schizophrenia is feasible. The effects and cost-effectiveness of music therapy for acute psychosis should be further explored in an explanatory randomised trial.

The Social Functioning Questionnaire: A Rapid and Robust Measure of Perceived Functioning.

BACKGROUND: The Social Functioning Questionnaire (SFQ), an eight-item selfreport scale (score range 0-24), was developed from the Social Functioning Schedule (SFS), a semi-structured interview which has been used primarily with non-psychotic patients and has good test-retest and inter-rater reliability as well as construct validity. The SFQ was developed following the need for a quick assessment of perceived social function. AIMS: To give further details of old and new data sets from studies involving over 4000 subjects assessed with the SFQ illustrating its epidemiological and clinical associations. METHOD: New data were analysed from a national epidemiological study, a comparison of key-worker and subject versions of the SFQ, and reanalysis of data from three earlier clinical studies, of psychiatric emergencies, general practice psychiatric patients and those with recurrent psychotic illnesses. These data were examined further to determine their range, their relationship to other clinical measures, and change over time in clinical trials. RESULTS: The population mean score in 4164 subjects was 4.6 and the data from all studies suggested that a score of 10 or more indicated poor social functioning. Those presenting as psychiatric emergencies had the poorest social function (mean 11.4) and psychiatric patients from general practice the best function (mean 7.7) of the clinical populations. The eight item scores had a normal distribution in psychiatric populations and a skewed one in a normal population; scores were relatively stable over the short (weeks) and long-term (months), and were high in the presence of acute mental health disturbance and personality disorder, giving support to the validity of the scale. The results from a UK sample of a randomly selected population specifically weighted for ethnic minorities showed similar social function across groups.

Social function, clinical symptoms and personality disturbance.

OBJECTIVES: To investigate the relationship between psychiatric symptoms, personality disturbance, and social function. METHOD: Longitudinal study of 100 psychiatric patients presenting as emergencies originally entered to a randomised trial of community and hospital-based treatment strategies. Ratings of social function using the Social Functioning Questionnaire, personality status using the Personality Assessment Schedule, and clinical symptomatology using the Comprehensive Psycho-pathological Rating Scale were recorded at baseline with assessment of social function repeated at two, four and 12 weeks. Correlation, regression, and path analysis were performed to test the hypothesis that personality status had more influence than clinical symptoms on social function. RESULTS: Path and regression analysis showed, that at baseline both psychopathology and personality pathology contributed to social dysfunction equally, but from two weeks onwards personality abnormality contributed to a greater degree than clinical psychopathology. Of the 100, 35 patients had a personality disorder and in these there was a strong correlation between social function scores at baseline and 12 weeks (48% of variation explained) whereas in those with no personality disorder the correlation was much weaker (14%); regression analyses confirmed this conclusion. CONCLUSIONS: Psychopathology and personality status contribute to social dysfunction in patients presenting as emergencies but persistent social dysfunction is more likely to reflect personality pathology than other forms of mental disorder.