Do the benefits of homeownership on mental health vary by race and poverty status? An application of doubly robust estimation for causal inference

While empirical studies have observed that homeownership is associated with improved mental health conditions, research indicates that this relationship might vary by race. Moreover, such a White-Black disparity in the impacts of homeownership on mental health could be complexed by poverty status, as maintaining one's homeownership could be a financial burden for people living in poverty status, defined by the US official poverty threshold. We add to the existing literature by analyzing the impacts of homeownership on psychological distress, simultaneously disaggregating by race and poverty status using survey data from the Panel Study on Income Dynamics from the 2017 and 2019 waves (N = 7059). Propensity score weighting and doubly robust estimation are applied to estimate causal inference for the impact of 2017 homeownership on 2019 psychological distress using negative binomial models. First, we found the impacts of homeownership on reducing psychological distress are significant for White Americans, not for Black Americans. Second, we found such a White-Black disparity is only observable for populations not living in poverty. On the other hand, for populations living in poverty, homeownership no longer lowers psychological distress for either race. Findings suggest that financial support and mental health support are needy to address inequality in the impacts of homeownership on mental health, which could simultaneously vary by poverty status and race. Implications are discussed.

Propensity Score Weighting with Missing Data on Covariates and Clustered Data Structure

Propensity score (PS) analyses are increasingly popular in behavioral sciences. Two issues often add complexities to PS analyses, including missing data in observed covariates and clustered data structure. In previous research, methods for conducting PS analyses with considering either issue alone were examined. In practice, the two issues often co-occur; but the performance of methods for PS analyses in the presence of both issues has not been evaluated previously. In this study, we consider PS weighting analysis when data are clustered and observed covariates have missing values. A simulation study is conducted to evaluate the performance of different missing data handling methods (complete-case, single-level imputation, or multilevel imputation) combined with different multilevel PS weighting methods (fixed- or random-effects PS models, inverse-propensity-weighting or the clustered weighting, weighted single-level or multilevel outcome models). The results suggest that the bias in average treatment effect estimation can be reduced, by better accounting for clustering in both the missing data handling stage (such as with the multilevel imputation) and the PS analysis stage (such as with the fixed-effects PS model, clustered weighting, and weighted multilevel outcome model). A real-data example is provided for illustration.

Covariate balance-related propensity score weighting in estimating overall hazard ratio with distributed survival data

BACKGROUND

When data is distributed across multiple sites, sharing information at the individual level among sites may be difficult. In these multi-site studies, propensity score model can be fitted with data within each site or data from all sites when using inverse probability-weighted Cox regression to estimate overall hazard ratio. However, when there is unknown heterogeneity of covariates in different sites, either approach may lead to potential bias or reduced efficiency. In this study, we proposed a method to estimate propensity score based on covariate balance-related criterion and estimate the overall hazard ratio while overcoming data sharing constraints across sites.

METHODS

The proposed propensity score was generated by choosing between global and local propensity score based on covariate balance-related criterion, combining the global propensity score fitted in the entire population and the local propensity score fitted within each site. We used this proposed propensity score to estimate overall hazard ratio of distributed survival data with multiple sites, while requiring only the summary-level information across sites. We conducted simulation studies to evaluate the performance of the proposed method. Besides, we applied the proposed method to real-world data to examine the effect of radiation therapy on time to death among breast cancer patients.

RESULTS

The simulation studies showed that the proposed method improved the performance in estimating overall hazard ratio comparing with global and local propensity score method, regardless of the number of sites and sample size in each site. Similar results were observed under both homogeneous and heterogeneous settings. Besides, the proposed method yielded identical results to the pooled individual-level data analysis. The real-world data analysis indicated that the proposed method was more likely to find a significant effect of radiation therapy on mortality compared to the global propensity score method and local propensity score method.

CONCLUSIONS

The proposed covariate balance-related propensity score in multi-site distributed survival data outperformed the global propensity score estimated using data from the entire population or the local propensity score estimated within each site in estimating the overall hazard ratio. The proposed approach can be performed without individual-level data transfer between sites and would yield the same results as the corresponding pooled individual-level data analysis.

Risk of post-extraction bleeding with direct oral anticoagulant compared with warfarin: Retrospective cohort study using large scale claims data in Japan

INTRODUCTION

Comparative safety of direct oral anticoagulants vs. warfarin in patients undergoing tooth extraction remains unclear. We compared the incidence of post-extraction bleeding between patients taking warfarin and those taking direct oral anticoagulants (DOACs) using administrative claims data.

MATERIALS AND METHODS

We identified outpatients on anticoagulant therapy who underwent permanent tooth extraction between 2015 and 2020 and categorized them into the warfarin and DOAC groups based on medication prescribed within six months prior to tooth extraction. We used the overlap propensity score weighting method to balance the baseline characteristics between the groups and compared the incidence of post-extraction bleeding within seven days after tooth extraction.

RESULTS

Among 5253 eligible patients, those in the DOAC group (n = 3696) were older and less frequently prescribed antiplatelets than those in the warfarin group (n = 1557). The distribution of tooth extraction type and number of teeth extracted in a single procedure did not differ between the groups. The unadjusted incidences of post-extraction bleeding in the warfarin and DOAC groups were 35 (2.2 %) and 71 (1.9 %), respectively. Moreover, the overlap weighting analysis showed that the adjusted odds ratio of post-extraction bleeding in the DOAC group in comparison with that in the warfarin group was 0.84 (95 % confidence interval, 0.54-1.31).

CONCLUSION

The incidence of post-extraction bleeding in patients taking DOACs was comparable to that in patients taking warfarin. The findings suggest that dentists and physicians should exercise the same degree of caution when extracting teeth in patients on DOACs and those on warfarin in terms of post-extraction bleeding.

Housing instability and mental health among renters in the michigan recession and recovery study

OBJECTIVE

To examine whether housing instability-inclusive of eviction, homelessness, moving in with others, moving for cost reasons, or frequent moves-is associated with mental health among renters in the aftermath of the Great Recession of 2007-09.

STUDY DESIGN

A panel survey study.

METHODS

We used data from the Michigan Recession and Recovery Study (2009-10, 2011 and 2013), a population-representative sample of working-aged adults, and logistic regression with propensity score weights to examine the association between housing instability over a year and a half and anxiety attack or depression symptoms at follow-up.

RESULTS

Respondents with any housing instability were 14 percentage points more likely to have had a recent anxiety attack, and those who had moved for cost reasons were 16 percentage points more likely. Respondents who experienced eviction were significantly more likely to meet criteria for major or minor depression at follow-up, by 13 percentage points.

CONCLUSIONS

Prior evidence of an association between housing instability and mental health is supported by these findings, which are robust to potential confounders, including financial and life shocks, housing quality, and neighborhood poverty concentration.

Methodological perspectives on the study of the health effects of unemployment - reviewing the mode of unemployment, the statistical analysis method and the role of confounding factors

Introduction

Studying the relationship between unemployment and health raises many methodological challenges. In the current study, the aim was to evaluate the sensitivity of estimates based on different ways of measuring unemployment and the choice of statistical model.

Methods

The Northern Swedish cohort was used, and two follow-up surveys thereof from 1995 and 2007, as well as register data about unemployment. Self-reported current unemployment, self-reported accumulated unemployment and register-based accumulated unemployment were used to measure unemployment and its effect on self-reported health was evaluated. Analyses were conducted with G-computation, logistic regression and three estimators for the inverse probability weighting propensity scores, and 11 potentially confounding variables were part of the analyses. Results were presented with absolute differences in the proportion with poor self-reported health between unemployed and employed individuals, except when logistic regression was used alone.

Results

Of the initial 1083 pupils in the cohort, our analyses vary between 488–693 individuals defined as employed and 61–214 individuals defined as unemployed. In the analyses, the deviation was large between the unemployment measures, with a difference of at least 2.5% in effect size when unemployed was compared with employed for the self-reported and register-based unemployment modes. The choice of statistical method only had a small influence on effect estimates and the deviation was in most cases lower than 1%. When models were compared based on the choice of potential confounders in the analytical model, the deviations were rarely above 0.6% when comparing models with 4 and 11 potential confounders. Our variable for health selection was the only one that strongly affected estimates when it was not part of the statistical model.

Conclusions

How unemployment is measured is highly important when the relationship between unemployment and health is estimated. However, misspecifications of the statistical model or choice of analytical method might not matter much for estimates except for the inclusion of a variable measuring health status before becoming unemployed. Our results can guide researchers when analysing similar research questions. Model diagnostics is commonly lacking in publications, but they remain very important for validation of analyses.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-022-01670-1.

Threshold heterogeneity of perioperative hemoglobin drop for acute kidney injury after noncardiac surgery: a propensity score weighting analysis

Background

Perioperative hemoglobin drop after noncardiac surgery is associated with acute kidney injury (AKI). However, opinion on the tolerable difference in postoperative hemoglobin drop in patients with different preoperative hemoglobin levels does not reach a consensus. This study aimed to identify hemoglobin drop thresholds for AKI after noncardiac surgery stratified by preoperative hemoglobin levels.

Method

This was a single-center retrospective cohort study for elective noncardiac surgery from January 1, 2012, to December 31, 2018. The endpoint was the occurrence of AKI 7 days postoperatively in the hospital. The generalized additive model described the non-linear relationship between hemoglobin drop and AKI occurrence. The minimum P-value approach identified cut-off points of hemoglobin drop within postoperative 7 days for patients with or without preoperative anemia. Stratified by preoperative anemia, hemoglobin drop’s odds ratio as continuous, quintile and dichotomous variables by various cut-off points for postoperative AKI were calculated in multivariate logistic regression models before and after propensity score weighting (PSW).

Results

Of the 35,631 surgery, 5.9% (2105 cases) suffered postoperative AKI. Non-linearity was found between hemoglobin drop and postoperative AKI occurrence. The thresholds and corresponding odds ratio of perioperative hemoglobin drop for patients with and without preoperative anemia were 18 g/L (1.38 (95%CI 1.14 -1.62), P < .001; after PSW: 1.42 (95%CI 1.17 -1.74), P < .001) and 43 g/L (1.81 (95%CI 1.35—2.27), P < .001; after PSW: 2.88 (95%CI 1.85—4.50), P < .001) respectively. Overall thresholds and corresponding odds ratio were 43 g/L (1.82 (95%CI 1.42—2.21)), P < .001; after PSW: 3.29 (95%CI 2.00—5.40), P < .001). Sensitivity analysis showed similar results. Heterogeneity subgroup analysis showed that intraoperatively female patients undergoing intraperitoneal surgery without colloid infusion seemed to be more vulnerable to higher hemoglobin drop. Further analysis showed a possible linear relationship between preoperative hemoglobin and perioperative hemoglobin drop thresholds. Additionally, this study found that the creatinine level changed simultaneously with hemoglobin level within five postoperative days.

Conclusions

Heterogeneity of hemoglobin drop endurability exists after noncardiac non-kidney surgery. More care and earlier intervention should be put on patients with preoperative anemia.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12882-022-02834-3.

Social distancing, trust and post-COVID-19 recovery

With the tourism and hospitality sector reopening post-lockdown of COVID-19, the recovery of customers' purchase intentions is essential to reboot the sector. This study aims to examine the relationship between social distancing measures and purchase intentions in the UK's restaurant and hotel sectors using a propensity score weighting experimental design method. The findings suggest that the impact of social distancing measures on purchase intentions is mediated by the trust in the targeted restaurant and hotel. Risk tolerance significantly moderates the influence of social distancing measures on trust; (non-) cash promotions have an insignificant impact on purchase intentions. The introduction of the propensity score weighting scheme addresses the endogeneity caused by the sampling bias in non-probability sampling experiment studies.

Childhood vaccination timeliness following maternal migration to an informal urban settlement in Kenya

INTRODUCTION

Timely receipt of recommended vaccines is a proven strategy to reduce preventable under-five deaths. Kenya has experienced impressive declines in child mortality from 111 to 43 deaths per 1000 live births between 1980 and 2019. However, considerable inequities in timely vaccination remain, which unnecessarily increases risk for serious illness and death. Maternal migration is a potentially important driver of timeliness inequities, as the social and financial stressors of moving to a new community may require a woman to delay her child's immunizations. This analysis examined how maternal migration to informal urban settlements in Nairobi, Kenya influenced childhood vaccination timeliness.

METHODS

Data came from the Nairobi Urban Health and Demographic Surveillance System, 2002-2018. Migration exposures were migrant status (migrant, non-migrant), migrant origin (rural, urban), and migrant type (first-time, circular [previously resided in settlement]). Age at vaccine receipt (vaccination timeliness) was calculated for all basic vaccinations. Accelerated failure time models were used to investigate relationships between migration exposures and vaccination timeliness. Confounding was addressed using propensity score weighting.

RESULTS

Over one-third of the children of both migrants and non-migrants received at least one dose late or not at all. Unweighted models showed the children of migrants had shorter time to OPV1 and DPT1 vaccine receipt compared to the children of non-migrants. After accounting for confounding only differences in timeliness for DPT1 remained, with the children of migrants receiving DPT1 significantly earlier than the children of non-migrants. Timeliness was comparable among migrants with rural and urban origins and among first-time and circular migrants.

CONCLUSION

Although a substantial proportion of children in Nairobi's informal urban settlements do not receive timely vaccination, this analysis found limited evidence that maternal migration and migration characteristics were associated with delays for most doses. Future research should seek to elucidate potential drivers of low vaccination timeliness in Kenya.

Gender differences in high-school dropout: Vulnerability and adolescent fertility in Chile

The original concerns about the consequences of adolescent fertility assumed that pregnancy is a turning point, which altered teens' life trajectories in terms of school progress, human capital accumulation and labor force participation, placing them on a path of vulnerability. However, several years of research have shown that teenagers who become pregnant are not a random sample of the population, but a selective sample, more likely to have limited socioeconomic resources and other characteristics that made them a vulnerable group to begin with. This paper studies the association between adolescent fertility and high school dropout in Chile taking that selectivity in consideration. We analyze the dropout of teen men and women, considering their socioeconomic status, sociodemographic characteristics, and characteristics of their sexual debut. Data comes the VIII Chilean Survey of the Youth, a nationally representative survey of people 15-29 years old applied in 2015. In order to deal with selectivity issues, we use a combination of propensity score weighting techniques and adjusted generalized linear models for estimating the effect of teen parenthood on high school dropout (ATT), for men and women separately. Our best estimates of the effect of teen parenting on the probability of high school dropout is 16-18 percent for women and ten percent for men, which implies that the educational setback of parenthood for women is about twice as high as the setback of men. These findings suggest the need of policies and interventions aimed both to reduce adolescent fertility, but also to facilitate the high school completion of those who already are parents.

The effect of racial discrimination on mental and physical health: A propensity score weighting approach

RATIONALE

The preponderance of research documents the negative consequences of racial discrimination for the mental and physical well-being across several racial/ethnic groups including Black Americans, Hispanic Americans, and Asian Americans. Despite this large body of research, few studies have attempted to evaluate racial discrimination as a casual factor of worse health among racial minority groups. The current study utilized nationally representative data to estimate the causal effect of racial discrimination on cardiovascular disease (CVD), self-rated physical health (SRH), body mass index (BMI), depression disorder, and substance use disorder. In addition, we examined whether the effect of racial discrimination on health was moderated by socio-demographic characteristics (i.e., age, race/ethnicity, sex, socioeconomic status, marital status, geographic location, and health insurance coverage).

METHODS

Data were from the combined National Survey of American Life and the National Latino and Asian American Study. We applied a propensity score weighting approach to estimate the differences between individuals who reported ever (N = 4358) or never (N = 1836) experiencing racial discrimination on a list of health outcomes (e.g., CVD, SRH, BMI, depression disorder, and substance use disorder).

RESULTS

Participants who reported ever experiencing racial discrimination were about 5 % higher CVD risk, had 0.12 points lower SRH, a 3 % higher probability of a depression disorder, and a 2 % higher probability of a substance use disorder. Moderation effects by race/ethnicity, sex, socioeconomic status, geographic region, and health insurance coverage were found.

CONCLUSION

Our study represents one of the first attempts to apply a propensity score weighting approach to causally link racial discrimination to worse health for racial minority individuals. This study adds to a larger body of research documenting the negative association between racial discrimination and health.

Evaluating treatments for posttraumatic stress disorder, alcohol and other drug use disorders using meta-analysis of individual patient data: Design and methodology of a virtual clinical trial

This paper describes Project Harmony, a Virtual Clinical Trial (VCT) funded by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) to harmonize and analyze data from over 40 independent psychological, pharmacologic and/or combined pharmacological treatment studies for posttraumatic stress disorder and comorbid alcohol and other drug use disorders (PTSD/AOD). The study attends to three distinct analysis challenges: (1) variation in measurement of PTSD/AOD across studies, time, populations and reporters, (2) cross-study variation in treatment effect sizes and (3) non-randomized, cross-study variation in the classification of treatments (despite within-study randomization of treatment arms). To address these challenges, the study combines meta-analysis of individual patient data (MIPD), integrative data analysis (IDA) and propensity score weighting (PSW) to integrate raw data from these clinical trials. This protocol shows how this VCT analytic framework was used to (1) develop commensurate scale scores of PTSD and AOD severity when measures vary across studies, (2) compare the efficacy of evidence-based treatment models for PTSD/AOD, (3) test for potential mediators of treatment effects on AOD and PTSD across treatment models, and (4) explore individual- and study-level moderators to inform for whom each of the treatment models works best. The advantages of the general VCT approach are juxtaposed against the limitations of single randomized controlled trials and conventional meta-analysis.

Pre- and postnatal maternal smoking and offspring smoking trajectories: Evidence from a 20-year birth cohort

Maternal smoking is associated with increased risk of smoking in the offspring. However, it remains unclear whether this association depends on the timing of exposure to maternal smoking. We investigated the association between prenatal and/or postnatal maternal smoking and offspring smoking during adolescence. Participants (N = 1661) were from the Québec Longitudinal Study of Child Development cohort. We identified longitudinal trajectories of maternal smoking from before pregnancy to child age 12 years using group-based trajectory modeling (GBTM). Adolescent (12-19 years) smoking trajectories were also identified using GBTM. Associations between maternal smoking and offspring smoking trajectories were estimated using multinomial logistic regressions. We used propensity score inverse probability weighting (IPW) to account for the differential distribution of maternal and familial characteristics across exposure groups. We identified four distinct groups for maternal smoking: no (66.1%), decreasing (5.6%), increasing (9.5%) and persistent (18.8%) smoking, and three adolescent smoking trajectories: abstinent, early-onset (before age 15) and late-onset (after age 15). In IPW-adjusted models, youth with mothers with decreasing, increasing and persistent smoking had higher risk of being early-onset smokers compared with youth with mothers in the non-smoking group. We also found that only youth whose mothers were persistent smokers had an increased risk of late-onset smoking. Regardless of timing, offspring exposure to maternal smoking is associated with increased risk of smoking during adolescence. More research is needed on how to create effective smoking cessation campaigns that span preconception, prenatal, and postnatal periods to help prevent intergenerational transmission of smoking behaviors.

Temporary employment, work stress and mental health before and after the Spanish economic recession

PURPOSE

This paper analyses the effects of temporary employment on work stress and mental health before (2006/2007) and during the economic recession (2011/2012), and examines whether the economic recession worsened these two health outcomes.

METHODS

To control for selection bias, propensity scores (PS) are computed separately for salaried men and women using microdata from two cross-sectional health surveys in Spain, considering temporary (treatment group) versus permanent employment (control group). Next, we use difference-in-difference estimators stratifying by age, education level, and regional unemployment differences using PS as weights.

RESULTS

Our results indicate that salaried worker with a temporary labour contract tends to have similar levels of high work stress and poorer mental health (only for men) than permanent ones for both periods. The economic recession does not appear to worsen both outcomes. However, when stratifying the sample, the economic recession is responsible for increasing stress among older temporary workers and male university graduates, without affecting women. Regarding mental health, we only find evidence of a negative impact of the economic recession on male temporary workers with university education.

CONCLUSION

The economic recession has not affected poor mental health, but it has accentuated the adverse effects of work stress among some specific subgroups of male workers. These subgroups require specific preventive and support actions as the Spanish firms have not substantially changed their hiring practices after the recession.

A state-wide quasi-experimental effectiveness study of the scale-up of school-wide Positive Behavioral Interventions and Supports

The three-tiered Positive Behavioral Interventions and Supports (PBIS) framework promotes the development of systems and data analysis to guide the selection and implementation of evidence-based practices across multiple tiers. The current study examined the effects of universal (tier 1) or school-wide PBIS (SW-PBIS) in one state's scale-up of this tier of the framework. Annual propensity score weights were generated to examine the longitudinal effects of SW-PBIS from 2006-07 through 2011-12. School-level archival and administrative data outcomes were examined using panel models with an autoregressive structure. The sample included 1316 elementary, middle, and high schools. Elementary schools trained in SW-PBIS demonstrated statistically significantly lower suspensions during the fourth and fifth study years (i.e., small effect size) and higher reading and math proficiency rates during the first two study years as well as in one and two later years (i.e., small to large effect sizes), respectively. Secondary schools implementing SW-PBIS had statistically significantly lower suspensions and truancy rates during the second study year and higher reading and math proficiency rates during the second and third study years. These findings demonstrate medium effect sizes for all outcomes except suspensions. Given the widespread use of SW-PBIS across nearly 26,000 schools in the U.S., this study has important implications for educational practices and policies.

The role of venues in structuring HIV, sexually transmitted infections, and risk networks among men who have sex with men

Background

Venues form part of the sampling frame for time-location sampling, an approach often used for HIV surveillance. While sampling location is often regarded as a nuisance factor, venues may play a central role in structuring risk networks. We investigated individual reports of risk behaviors and infections among men who have sex with men (MSM) attending different venues to examine structuring of HIV risk behaviors. However, teasing apart ‘risky people’ from ‘risky places’ is difficult, as individuals cannot be randomized to attend different venues. However, we can emulate this statistically using marginal structural models, which inversely weight individuals according to their estimated probability of attending the venue.

Methods

We conducted a cross-sectional survey of 609 MSM patrons of 14 bars in San Diego, California, recruited using the Priorities for Local AIDS Control Efforts (PLACE) methodology, which consists of a multi-level identification and assessment of venues for HIV risk through population surveys.

Results and discussion

Venues differed by many factors, including participants’ reported age, ethnicity, number of lifetime male partners, past sexually transmitted infection (STI), and HIV status. In multivariable marginal structural models, venues demonstrated structuring of HIV+ status, past STI, and methamphetamine use, independently of individual-level characteristics.

Conclusions

Studies using time-location sampling should consider venue as an important covariate, and the use of marginal structural models may help to identify risky venues. This may assist in widespread, economically feasible and sustainable targeted surveillance and prevention. A more mechanistic understanding of how ‘risky venues’ emerge and structure risk is needed.

Electronic supplementary material

The online version of this article (10.1186/s12889-018-5140-3) contains supplementary material, which is available to authorized users.

Measuring multiple discrimination through a survey-based methodology

This paper focuses on the concept of multiple discrimination and its measurement through survey methods. The study was designed as a quasi-experimental comparison of survey mode effects on the quality of discrimination measurement: the traditional 'face-to-face' survey, the conventional self-completed mode and CAWI (finally deleted due to its non-comparability). Consistent with our hypothesis, some support was obtained for the social desirability bias and survey mode effects: 1) self-administration of questionnaires favours the declaration of discriminatory attitudes and personal experiences of discrimination; 2) the effect of privacy is greater in direct indicators of discriminatory attitudes; 3) perceptions and experiences of discrimination are more frequently reported by highly educated respondents. Nevertheless, contrary to our expectations, less educated respondents are also affected by survey mode and continue to be underrepresented in self-completed methods. The current research aims to serve as a basis for further research in this area.

The survival difference between gastric cancer patients from the UK and Japan remains after weighted propensity score analysis considering all background factors

BACKGROUND

Previous studies comparing survival between gastric cancer (GC) patients from the West and the East were based on the assumption that background factors and prognostic factors were identical. The aim of the current study was to compare the survival of GC patients from the UK and Japan using weighted propensity score analysis after identifying all different background factors.

METHODS

Data from 464 patients from the Leeds Teaching Hospital NHS Trust, Leeds, UK (LTHT), and 465 patients from the Kanagawa Cancer Center Hospital, Yokohama, Japan (KCCH), who had surgery for GC were analyzed. Prognostic factors for overall survival (OS) and cancer-specific survival (CSS) were identified by univariate and multivariate analyses. Survival was compared by propensity score weighting after adjusting for all significantly different background factors.

RESULTS

Most background factors were different between LTHT and KCCH patients. Unadjusted stage-specific OS and CSS were significantly better in KCCH. Independent prognostic factors for unadjusted OS and CSS were pT and pN in KCCH and in addition tumor location, pancreatectomy, resection margin status and number of examined lymph nodes in LTHT. Even after adjusting for all background characteristics, survival remained better in KCCH.

CONCLUSIONS

These results suggest that differences in background factors are unable to fully explain the survival difference of GC patients between UK and Japan. Comprehensive studies into the biology of GC and/or host factors are needed to fully understand the survival difference.