Methods This paper reviews fundamental LQAS design principles, and compares and contrasts the two LQAS methods. We raise four concerns with the simply-written, freely-downloadable training materials associated with the second method.

Results The first method is founded on sound statistical principles and is carefully designed to protect the vulnerable populations that it studies. The language used in the training materials for the second method is simple, but not at all clear, so the second method sounds very much like the first. On close inspection, however, the second method is found to promote study designs that are biased in favor of finding programmatic or intervention success, and therefore biased against the interests of the population being studied.

Conclusion We outline several recommendations, and issue a call for a new high standard of clarity and face validity for those who design, conduct, and report LQAS studies.

Methods The Mozambique Ministry of Health conducted a Malaria Indicator Survey (MIS) in June and July 2007. We applied LQAS classification rules to the 345 sampled enumeration areas to demonstrate identifying high- and low-performing areas with respect to two malaria program indicators—‘household possession of any bednet’ and ‘household possession of any insecticide-treated bednet (ITN)’.

Results As shown by the MIS, no province in Mozambique achieved the 70% coverage target for household possession of bednets or ITNs. By applying LQAS classification rules to the data, we identify 266 of the 345 enumeration areas as having bednet coverage severely below the 70% target. An additional 73 were identified with low ITN coverage.

Conclusions This article demonstrates the feasibility of integrating LQAS into multistage cluster sampling surveys and using these results to support a comprehensive national, regional and local programme M&E system. Furthermore, in the recommendations we outlined how to integrate the Large Country-LQAS design into macro-surveys while still obtaining results available through current sampling practices.

Methods This article presents a method for correcting the bias in FR estimates from such misclassification of family history, using a simple formula that depends on the prevalence and sensitivity of the observed family history. The sensitivity is estimated by using the R package poplab to create realistic populations of related individuals and then imposing the start-up effect of disease registration.

Results For a range of FRs, the truncation of family history is demonstrated to result in non-differential misclassification, and sensitivity that has little or no dependence on the FR. The bias is most pronounced for high FRs and for registers with a short life span, and increases with the age of the study cohort. In all the situations studied, the bias-corrected estimates are in excellent agreement with the true values.

Conclusions In summary, our method can correct the inevitable bias in FRs induced by using electronic population data, and is a feasible alternative to the use of validation samples.

Methods Methodological review to address the concept of external validity.

Results External validity refers to the question whether results are generalizable to persons other than the population in the original study. The only formal way to establish the external validity would be to repeat the study for that specific target population. We propose a three-way approach for assessing the external validity for specified target populations. (i) The study population might not be representative for the eligibility criteria that were intended. It should be addressed whether the study population differs from the intended source population with respect to characteristics that influence outcome. (ii) The target population will, by definition, differ from the study population with respect to geographical, temporal and ethnical conditions. Pondering external validity means asking the question whether these differences may influence study results. (iii) It should be assessed whether the study's conclusions can be generalized to target populations that do not meet all the eligibility criteria.

Conclusion Judging the external validity of study results cannot be done by applying given eligibility criteria to a single target population. Rather, it is a complex reflection in which prior knowledge, statistical considerations, biological plausibility and eligibility criteria all have place.

Methods Each method is based on assumed associations between confounder and exposure, confounder and outcome and the prevalence of the confounder in the population at large. In the first method an unmeasured confounder is simulated and subsequently adjusted. The other two methods are analytical methods, in which either the (adjusted) effect estimate is multiplied with a factor based on assumed confounder characteristics, or the (adjusted) risks for the outcome among exposed and unexposed subjects are adjusted by such a factor. These methods are illustrated with a clinical example on influenza vaccine effectiveness.

Results When applied to a dataset constructed to assess the effect of influenza vaccination on mortality, the three reviewed methods provided similar results. After adjustment for observed confounders, influenza vaccination reduced mortality by 42% [odds ratio (OR) 0.58, 95% confidence interval (CI) 0.46–0.73]. To arrive at a 95% CI including one requires a very common confounder (40% prevalence) with strong associations with both vaccination status and mortality, respectively OR ≤0.3 and OR ≥3.0 (OR 0.79, 95% CI 0.62–1.00).

Conclusions In every non-randomized study on causal associations the robustness of the results with respect to unmeasured confounding can, and should, be assessed using sensitivity analyses.

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.

Methods In the baseline examination of ESTHER, a study conducted in Saarland, serological measurements of pepsinogen I and II and H. pylori antibodies were taken in 9444 participants aged 50–74 years. Information on potential risk factors and medical history were obtained by self-administered standardized questionnaire.

Results A strong dose–response relationship between sibship size and H. pylori seroprevalence was observed (P < 0.01). Adjusted odds ratios (ORs) 95% confidence interval (CI) for H. pylori seropositivity for subjects with 4, 5, 6 and 7 or more siblings compared with subjects without siblings were 1.45 (1.20–1.77), 1.83 (1.50–2.22) and 1.84 (1.47–2.31), respectively. A large sibship size was also associated with an increased risk of CAG with an adjusted OR of 1.42 (1.01–2.01) for 7 or more compared with less than or equal to 2 siblings. This association was attenuated but not entirely eliminated after additional adjustment for H. pylori infection. Notably, a significant association between large sibship size and CAG was also found among H. pylori-negative subjects.

Conclusions Our results suggest that large sibship size is associated with increased H. pylori prevalence and CAG risk. The association with CAG risk may be mediated at least in part by H. pylori infection. However, mechanisms other than H. pylori infection may contribute to the ‘sibling effect’ as well.

Methods In the French nationwide Mortalité 2000 and 2005 surveys, physicians reported causes of deaths in HIV-infected adults in 2000 and 2005, using a standardized questionnaire. We used multivariate logistic regression models to study the association between gender and AIDS-defining causes of death, adjusting for other characteristics.

Results Of the 1013 HIV-infected adults who died in 2005, 247 (24%) were women. Half of women were infected through heterosexual contacts, compared with 25% men. In 2005, the proportion of AIDS-defining causes of death was higher in women than in men (43 vs 34%; P = 0.01), whereas it had been the same in 2000 (47% in women and men). In 2005, women died less frequently than men from respiratory malignancies (lung, ear/nose/throat) and cardiovascular disease (9% of all causes of death in women compared with 16% in men; P = 0.004), and suicides or accidents (4 vs 9%; P = 0.02). Socio-economic precariousness, younger age, less alcohol and tobacco consumption and lack of prior ART explained the higher proportion of deaths from AIDS in women compared with men.

Conclusions The higher proportion of AIDS-related deaths in women is probably explained by two factors: (i) some HIV-infected women, especially migrants in poor socio-economic conditions, may not have access to optimal care; and (ii) a lower prevalence of risk factors for respiratory, cardiovascular and violent deaths means that the risk of dying from non-AIDS causes may be lower in women.

Methods A systematic literature review was carried out to identify cohort studies that collected data on weight and height at baseline and that used a diagnosis of active TB as the study outcome. Weight-for-height measures used in the original studies were transformed into body mass index (BMI). Exponential trend lines were fitted to each data set.

Results Six studies were included. In all of them, there was a log-linear inverse relationship between TB incidence and BMI, within the BMI range 18.5–30 kg/m². The average slope gave a reduction in TB incidence of 13.8% [95% confidence interval 13.4–14.2] per unit increase in BMI. The dose–response relationship was less certain at BMI <18.5 and >30 kg/m².

Conclusion There is a strong and consistent log-linear relationship between TB incidence and BMI across a variety of settings with different levels of TB burden. More research is required to test the relationship at very low and very high BMI levels, to establish the biological mechanism linking BMI with risk of TB and to establish the potential impact on the global TB epidemic of changing nutritional status of populations.

Methods The 2323 cases were categorized as ‘ever screened’ if they had had at least one smear prior to 6 months before diagnosis, and as ‘regular screening’ if they had had no more than 36 months between any two smears in the period 6–114 months before diagnosis. Logistic regression was used to estimate the associations of screening history, ethnicity, SES and urban/rural residence with stage at diagnosis.

Results The percentages ‘ever screened’ were 43.3% overall, 24.8% in Pacific, 30.5% in Asian, 40.6% in Maori and 46.1% in ‘Other’ women. The corresponding estimates for ‘regular screening’ were 14.0, 5.7, 7.8, 12.5 and 15.3%. Women with ‘regular screening’ had a lower risk of late stage diagnosis [odds ratio (OR) 0.16, 95% confidence interval (CI) 0.10–0.26], and the effect was greater for squamous cell carcinoma (OR 0.12, 95% CI 0.07–0.23) than for adenocarcinoma (OR 0.32, 95% CI 0.13–0.82). The increased risk of late-stage diagnosis (OR 2.72, 95% CI 1.99–3.72) in Maori (compared with ‘Other’) women decreased only slightly when adjusted for screening history (OR 2.45, 95% CI 1.77–3.39).

Conclusions Over half of cases had not been ‘ever screened’. Regular screening substantially lowered the risk of being diagnosed at a late stage. However, screening history does not appear to explain the ethnic differences in stage at diagnosis.

Methods We undertook a pooled analysis of four population-based and four hospital-based case–control studies from the International Head and Neck Cancer Epidemiology (INHANCE) consortium, with participants from Argentina, Australia, Brazil, Canada, Cuba, India, Italy, Spain, Poland, Puerto Rico, Russia and the USA. The study included 5642 head and neck cancer cases and 6069 controls. We calculated odds ratios (ORs) of associations between cancer and specific sexual behaviours, including practice of oral sex, number of lifetime sexual partners and oral sex partners, age at sexual debut, a history of same-sex contact and a history of oral–anal contact. Findings were stratified by sex and disease subsite.

Results Cancer of the oropharynx was associated with having a history of six or more lifetime sexual partners [OR = 1.25, 95% confidence interval (CI) 1.01, 1.54] and four or more lifetime oral sex partners (OR = 2.25, 95% CI 1.42, 3.58). Cancer of the tonsil was associated with four or more lifetime oral sex partners (OR = 3.36, 95 % CI 1.32, 8.53), and, among men, with ever having oral sex (OR = 1.59, 95% CI 1.09, 2.33) and with an earlier age at sexual debut (OR = 2.36, 95% CI 1.37, 5.05). Cancer of the base of the tongue was associated with ever having oral sex among women (OR = 4.32, 95% CI 1.06, 17.6), having two sexual partners in comparison with only one (OR = 2.02, 95% CI 1.19, 3.46) and, among men, with a history of same-sex sexual contact (OR = 8.89, 95% CI 2.14, 36.8).

Conclusions Sexual behaviours are associated with cancer risk at the head and neck cancer subsites that have previously been associated with HPV infection.

Methods We pooled individual-level data from case–control studies in the International Head and Neck Cancer Epidemiology Consortium. Data were available from 13 studies on drinking cessation (9167 cases and 12 593 controls), and from 17 studies on smoking cessation (12 040 cases and 16 884 controls). We estimated the effect of quitting smoking and drinking on the risk of head and neck cancer and its subsites, by calculating odds ratios (ORs) using logistic regression models.

Results Quitting tobacco smoking for 1–4 years resulted in a head and neck cancer risk reduction [OR 0.70, confidence interval (CI) 0.61–0.81 compared with current smoking], with the risk reduction due to smoking cessation after ≥20 years (OR 0.23, CI 0.18–0.31), reaching the level of never smokers. For alcohol use, a beneficial effect on the risk of head and neck cancer was only observed after ≥20 years of quitting (OR 0.60, CI 0.40–0.89 compared with current drinking), reaching the level of never drinkers.

Conclusions Our results support that cessation of tobacco smoking and cessation of alcohol drinking protect against the development of head and neck cancer.

Methods This was a prospective cohort study. Participants consisted of 13 621 men and women from the Aerobics Center Longitudinal Study (ACLS). Participants completed a clinical exam and 3-day diet record between 1987 and 1999. Participants were followed for mortality until 2003. Reduced rank regression (RRR) was used to identify dietary patterns that predicted unfavourable total and high-density lipoprotein-cholesterol, triglyceride, glucose, blood pressure, uric acid, white blood cell and body mass index values.

Results One primary dietary pattern emerged and was labelled the Unhealthy Eating Index. This pattern was characterized by elevated consumption of processed and red meat, white potato products, non-whole grains, added fat and reduced consumption of non-citrus fruits. The hazard ratio for all-cause mortality in the fifth vs the first quintile of the Unhealthy Eating Index was 1.40 (1.02–1.91). This risk estimate was reduced by 13.5 and 55.0% after controlling for self-reported physical activity and fitness, respectively.

Conclusion In this study the association between diet and overall mortality was, in large part, confounded by fitness.

Methods A follow-up study was conducted for 16 828 men and 8368 women aged 30–59 years and apparently healthy at baseline. Incident DM was identified by ‘fasting serum glucose ≥7.00 mmol/l (126 mg/dl)’ or/and ‘under medical treatment for DM’. Using Cox proportional hazard models, hazard ratio (HR) for incident DM were estimated according to the quartiles of heart rate, systolic or diastolic BP (SBP, DBP), and its linear trends were checked by computing the three indices as continuous variables. Subsequently, interactive effects of slow/fast heart rate (dichotomized by the median) and low/high SBP or DBP (dichotomized by the median) on HR were examined. Baseline age, body mass index, smoking, drinking, exercise and education were computed as conventional confounders.

Results During the follow-up of 125 106 person-years for men and 59 616 person-years for women, 869 men and 224 women developed DM. The multivariate-adjusted HR for incident DM increased across quartiles of heart rate, SBP and DBP in both sexes (linear trend P<0.001 for all). Neither sex showed any significant interactive effects of heart rate and SBP or DBP on HR.

Conclusions Resting heart rate and BP proportionally raise the risk for DM in middle-aged healthy men and women. Moreover, the adverse effects of fast heart rate and high BP are independent of each other as well as of the influences of conventional confounders.

Methods A total of 6640 men and women, aged 25–84 years, were included in a 10-year mortality follow-up: 345 subjects with a diagnosed AAA and 6295 subjects with a non-aneurismal aorta. Non-aneurismal aortic diameter and prevalent AAAs were categorized into seven groups.

Results In subjects without an AAA, an aortic diameter ≥30 mm increased age- and sex-adjusted total mortality [mortality rate ratio (MRR) = 3.73, 95% confidence interval (CI) 1.77–7.89] and CVD mortality (MRR = 9.24, 95% CI 4.07–20.97) compared with subjects with aortic diameter of 21–23 mm. An AAA at screening was strongly associated with deaths from aortic aneurysm and was associated with total (MRR = 1.60, 95% CI 1.31–1.96) and CVD mortality (MRR = 2.41, 95% CI 1.81–3.21). This was not explained by deaths due to an AAA. Adjustments for CVD risk factors could fully explain the increased total, but not CVD mortality in subjects with an AAA.

Conclusions An AAA increases total and CVD mortality. In the large majority of subjects with a non-aneurysmal aorta, the diameter does not influence total or CVD mortality. However, in individuals with a maximal diameter >26 mm (2% of the population), a positive relationship is found.

Methods From 1988 to 1990, a total of 67 129 Japanese non-workers or daytime workers (27 755 men and 39 374 women) aged 40–79 years, without a history of stroke, heart disease or cancer, completed a lifestyle questionnaire. They were followed for mortality until the end of 2003.

Results During the 879 244 person-year follow-up, 9643 deaths (2852 from CVD, 3643 from cancer, 2392 from other internal causes, 738 from external causes and 18 from unspecified causes) were observed. After adjustment for possible confounders, subjects with a daytime napping habit had elevated hazard ratios (HRs) for mortality from all causes [HR 1.19, 95% confidence interval (CI) 1.14–1.24, P < 0.001], CVD (HR 1.31, 95% CI 1.22–1.42, P < 0.001), non-cardiovascular/non-cancer internal diseases (HR 1.26, 95% CI 1.16–1.37, P < 0.001) and external causes (HR 1.28, 95% CI 1.10–1.50, P = 0.001), but not for cancer death (HR 1.03, 95% CI 0.96–1.10, P = 0.400). The risk of CVD mortality associated with daytime napping was diminished among overweight subjects, but pronounced in those with weight loss after age 20 years, with non-regular employment, with lower education level and with a follow-up period <5 years.

Conclusions Daytime napping is associated with elevated risk of CVD mortality as well as non-cardiovascular/non-cancer and external deaths. Daytime napping may elevate risk of CVD death through some biological effects but, to a larger extent, some comorbid disorders causing weight loss or associated with non-regular employment and low education level could explain this association.

Methods A cross-sectional survey with face-to-face interviews of rural and urban residents in northern China. Before providing any information, we asked the residents whether they would accept drug treatment if they had hypertension and also asked them to estimate the 5-year cardiovascular risk in untreated hypertension and the benefits from anti-hypertensive drugs. We then informed the participants of necessary information and asked them above what benefit they would be willing to pay the current cost, and how much they would be willing to pay for the actual benefit, for anti-hypertensive drugs out of pocket.

Results Eight hundred and eighty-seven rural residents and 921 urban residents were interviewed with a response rate of 97%. Ninety-five percent [95% confidence interval (CI) 94–96%] of the residents said they would take anti-hypertensive drugs if they had hypertension, although 91% (95% CI 89–92%) said they did not have sufficient knowledge to make a decision. Seventy-eight percent (95% CI 76–80%) believed that anti-hypertensive drugs were primarily to lower blood pressure or relieve symptoms. They over-estimated the cardiovascular risk of untreated hypertension by approximately 12 times and the absolute benefit of drug treatment by 20 times. Given the actual absolute benefit of the drugs, only 23% (95% CI 21–25%) were willing to pay the current annual cost of $500 Ren Min Bi (US$73.3, 54.8 as of 8 May 2009) for these drugs. Given the current cost, they were, on average, willing to pay for the drugs only when the 5-year cardiovascular disease risk was as high as 35% (95% CI 31–38%) or even higher.

Conclusion The public in China are significantly misinformed and considerably over estimate the risk of hypertension and the benefit of treatment. The public's willingness to pay for anti-hypertensive drugs is much lower than the current guidelines implicitly assume. The willingness to pay should be considered, along with other factors, when prescribing anti-hypertensive drugs to an individual patient or making hypertension guidelines for a population.

Methods Proportional hazards Cox models were fitted to estimate rate ratios (RRs) for the separate and joint effects of entry and follow-up Coulter variables.

Results Entry RBC, PCV and Hgb were significant risk factors for CHD mortality after adjustment for gender but only PCV remained significant after adjustment for potential confounders [RR per 1 standard deviation (SD) increase = 1.17, 95% confidence interval (CI) 1.00–1.37]. This effect was partly reduced when the values of 6 years were analysed (RR per 1 SD increase = 1.10, 95% CI 0.93–1.30). No significant gender, smoking or age/time interactions were identified. PCV was the only significant predictor when all Coulter variables were studied jointly.

Conclusion PCV was found to predict CHD mortality even after controlling for classical risk factors. This may give some insight into possible mechanisms, such as an influence on thrombin production.

Methods Spatial accessibility, as measured by network travel times, to 457 grocery stores located in 205 neighbourhoods in four environmental settings (island, rural, small town and urban) in Scotland was calculated using Geographical Information Systems. The distribution of accessibility by neighbourhood deprivation in each of these four settings was investigated.

Results Overall, the most deprived neighbourhoods had the best access to grocery stores and grocery stores selling fresh produce. Stratified analysis by environmental setting suggests that the least deprived compared with the most deprived urban neighbourhoods have greater accessibility to grocery stores than their counterparts in island, rural and small town locations. Access to fresh produce is better in more deprived compared with less deprived urban and small town neighbourhoods, but poorest in the most affluent island communities with mixed results for rural settings.

Conclusions The results presented here suggest that the assumption of a universal ‘deprivation amplification’ hypothesis in studies of the neighbourhood food environment may be misguided. Associations between neighbourhood deprivation and grocery store accessibility vary by environmental setting. Theories and policies aimed at understanding and rectifying spatial inequalities in the distribution of neighbourhood exposures for poor diet need to be context specific.

Methods All children born in Pelotas, Brazil, in 2004, were recruited to a birth cohort study. These children were assessed at birth and at 3, 12 and 24 months of age. In this last assessment involving 3869 children, detailed information on socio-economic and health characteristics was collected. Child development was assessed using the screening version of Battelle's Development Inventory. Five markers of cognitive stimulation and social interaction were recorded and summed to form a score ranging from 0–5. The outcomes studied were mean development score and low performance (less than 10th percentile of the sample).

Results Child development was strongly associated with socio-economic position, maternal schooling and stimulation. Having been told a story and owning a book were the least frequent markers among children with score 1. These children were 8.3 times more likely to present low performance than those who scored 5. The effect of stimulation was much stronger among children from mothers with a low level of schooling—one additional point added 1.7 on the child's development for children of low-schooling mothers, whereas only 0.6 was added for children of high-schooling mothers.

Conclusions Our stimulation markers cannot be directly translated into intervention strategies, but strongly suggest that suitably designed cognitive stimulation can have an important effect on children, especially those from mothers with low schooling.

Methods We used multivariable linear regression to examine the prospective adjusted associations of breastfeeding with body mass index (BMI), height and weight z-scores at 7 years of age relative to the 2007 World Health Organization (WHO) growth reference, in a large (n = 8327), population-representative Hong Kong Chinese birth cohort, recruited between April and May 1997 with high follow-up (n = 7026).

Results Low socio-economic position (SEP) was associated with never breastfeeding and with exclusive breastfeeding for ≥3 months. We did not find any association between breastfeeding and BMI [z-score mean difference 0.07, 95% confidence interval (CI) –0.05 to 0.19], height (0.02, 95% CI –0.07 to 0.11) or weight (0.07, 95% CI –0.05 to 0.18), adjusted for sex, birth weight, gestational age, SEP, second-hand smoke (SHS) exposure, parity, mother's age at birth, mother's place of birth and serious infant morbidity.

Conclusions In a non-European setting, breastfeeding was not associated with child adiposity, suggesting that observed protective effects may be due to socially patterned confounding by SEP, maternal adiposity and maternal smoking.