Multiple imputation to minimise bias from missing stage information in estimates of early cancer diagnosis in England: a population-based study

IntroductionMonitoring early diagnosis is a priority of cancer policy in England. Information on stage has not always been available for a large proportion of patients, however, which may bias temporal comparisons. We previously estimated that early-stage diagnosis of colorectal cancer rose from 32% to 44% during 2008–2013, using multiple imputation. Here we examine the underlying assumptions of multiple imputation for missing stage using the same dataset.MethodsIndividually-linked cancer registration, Hospital Episode Statistics (HES), and audit data were examined. Six imputation models including different interaction terms, post-diagnosis treatment, and survival information were assessed, and comparisons drawn with the a priori optimal model. Models were further tested by setting stage values to missing for some patients under one plausible mechanism, then comparing actual and imputed stage distributions for these patients. Finally, a pattern-mixture sensitivity analysis was conducted.ResultsData from 196,511 colorectal patients were analysed, with 39.2% missing stage. Inclusion of survival time increased the accuracy of imputation: the odds ratio for change in early-stage diagnosis during 2008–2013 was 1.7 (95% CI: 1.6, 1.7) with survival to 1 year included, compared to 1.9 (95% CI 1.9–2.0) with no survival information. Imputation estimates of stage were accurate in one plausible simulation. Pattern-mixture analyses indicated our previous analysis conclusions would only change materially if stage were misclassified for 20% of the patients who had it categorised as late.InterpretationMultiple imputation models can substantially reduce bias from missing stage, but data on patient’s one-year survival should be included for highest accuracy.     

Imputation of missing prostate cancer stage in English cancer registry data based on clinical assumptions

BackgroundCancer stage can be missing in national cancer registry records. We explored whether missing prostate cancer stage can be imputed using specific clinical assumptions.MethodsProstate cancer patients diagnosed between 2010 and 2013 were identified in English cancer registry data and linked to administrative hospital and mortality data (n = 139,807). Missing staging items were imputed based on specific assumptions: men with recorded N-stage but missing M-stage have no distant metastases (M0); low/intermediate-risk men with missing N- and/or M-stage have no nodal disease (N0) or metastases; and high-risk men with missing M-stage have no metastases. We tested these clinical assumptions by comparing 4-year survival in men with the same recorded and imputed cancer stage. Multi-variable Cox regression was used to test the validity of the clinical assumptions and multiple imputation.ResultsSurvival was similar for men with recorded N-stage but missing M-stage and corresponding men with M0 (89.5% vs 89.6%); for low/intermediate-risk men with missing M-stage and corresponding men with M0 (92.0% vs 93.1%); and for low/intermediate-risk men with missing N-stage and corresponding men with N0 (90.9% vs 93.7%). However, survival was different for high-risk men with missing M-stage and corresponding men with M0. Imputation based on clinical imputation performs as well as statistical multiple imputation.ConclusionSpecific clinical assumptions can be used to impute missing information on nodal involvement and distant metastases in some patients with prostate cancer.     

Approaches for missing covariate data in logistic regression with MNAR sensitivity analyses

Data with missing covariate values but fully observed binary outcomes are an important subset of the missing data challenge. Common approaches are complete case analysis (CCA) and multiple imputation (MI). While CCA relies on missing completely at random (MCAR), MI usually relies on a missing at random (MAR) assumption to produce unbiased results. For MI involving logistic regression models, it is also important to consider several missing not at random (MNAR) conditions under which CCA is asymptotically unbiased and, as we show, MI is also valid in some cases. We use a data application and simulation study to compare the performance of several machine learning and parametric MI methods under a fully conditional specification framework (MI-FCS). Our simulation includes five scenarios involving MCAR, MAR, and MNAR under predictable and nonpredictable conditions, where “predictable” indicates missingness is not associated with the outcome. We build on previous results in the literature to show MI and CCA can both produce unbiased results under more conditions than some analysts may realize. When both approaches were valid, we found that MI-FCS was at least as good as CCA in terms of estimated bias and coverage, and was superior when missingness involved a categorical covariate. We also demonstrate how MNAR sensitivity analysis can build confidence that unbiased results were obtained, including under MNAR-predictable, when CCA and MI are both valid. Since the missingness mechanism cannot be identified from observed data, investigators should compare results from MI and CCA when both are plausibly valid, followed by MNAR sensitivity analysis.     

Multiple imputation methods for longitudinal blood pressure measurements from the Framingham Heart Study

Missing data are a great concern in longitudinal studies, because few subjects will have complete data and missingness could be an indicator of an adverse outcome. Analyses that exclude potentially informative observations due to missing data can be inefficient or biased. To assess the extent of these problems in the context of genetic analyses, we compared case-wise deletion to two multiple imputation methods available in the popular SAS package, the propensity score and regression methods. For both the real and simulated data sets, the propensity score and regression methods produced results similar to case-wise deletion. However, for the simulated data, the estimates of heritability for case-wise deletion and the two multiple imputation methods were much lower than for the complete data. This suggests that if missingness patterns are correlated within families, then imputation methods that do not allow this correlation can yield biased results.     

Missing Oral Health-Related Data in the interRAI-HC - Associations with Selected Variables of General Health and the Effect of Multiple Imputation on the Relationship between Oral and General Health

Background

Missing data within the comprehensive geriatric assessment of the interRAI suite of assessment instruments potentially imply the under-detection of conditions that require care as well as the risk of biased statistical results. Impaired oral health in older individuals has to be registered accurately as it causes pain and discomfort and is related to the general health status.

Objective

This study was based on interRAI-Home Care (HC) baseline data from 7590 subjects (mean age 81.2 years, SD 6.9) in Belgium. It was investigated if missingness of the oral health-related items was associated with selected variables of general health. It was also determined if multiple imputation of missing data affected the associations between oral and general health.

Materials and Methods

Multivariable logistic regression was used to determine if the prevalence of missingness in the oral health-related variables was associated with activities of daily life (ADLH), cognitive performance (CPS2) and depression (DRS). Associations between oral health and ADLH, CPS2 and DRS were determined, with missing data treated by 1. the complete-case technique and 2. by multiple imputation, and results were compared.

Results

The individual oral health-related variables had a similar proportion of missing values, ranging from 16.3% to 17.2%. The prevalence of missing data in all oral health-related variables was significantly associated with symptoms of depression (dental prosthesis use OR 1.66, CI 1.41–1.95; damaged teeth OR 1.74, CI 1.48–2.04; chewing problems OR 1.74, CI 1.47–2.05; dry mouth OR 1.65, CI 1.40–1.94). Missingness in damaged teeth (OR 1.27, CI 1.08–1.48), chewing problems (OR 1.22, CI 1.04–1.44) and dry mouth (OR 1.23, CI 1.05–1.44) occurred more frequently in cognitively impaired subjects. ADLH was not associated with the prevalence of missing data. When comparing the complete-case technique with the multiple imputation approach, nearly identical odds ratios characterized the associations between oral and general health.

Conclusion

Cognitively impaired and depressive individuals had a higher risk of missing oral health-related information. Associations between oral health and ADLH, CPS2 and DRS were not influenced by multiple imputation of missing data. Further research should concentrate on the mechanisms that mediate the occurrence of missingness to develop preventative strategies.     

Difference in net survival using regional and national life tables in Japan

BackgroundNational life table is commonly used for estimating cancer net survival. However, the national life table does not reflect condition of people in local area accurately, because there are disparities in cancer mortality rates among the local area in many cases. We investigated magnitude of difference in cancer net survival using the local area in Japan and Japanese life tables.MethodsWe analyzed data from 32,942 cancer patients diagnosed between 2010 and 2012 in Aomori prefecture, Japan. Expected survival rates in Aomori (ESA) and Japan (ESJ) were estimated based on the life table of each area. Five-year net survival rates using ESA and the ESJ were estimated using the Pohar-Perme method.ResultsThe difference between net survival rates using the ESA (NSA) and the ESJ (NSJ) were larger than in men (0.3－3.0%) than in women (0.1－0.8%). The largest difference in the net survival rate was observed in prostate cancer patients, because the difference in the expected survival in oldest old men was remarkable.ConclusionTwo factors affected the difference in the net survival rates resulting from the sensitivity analysis. The difference was larger (1) among older patients or (2) with a longer observation period (person-years).     

Trends in age-standardised net survival of stomach cancer by subsite and stage: A population-based study in Osaka,Japan, 2001-2014

IntroductionThe burden of stomach cancer remains high, particularly among Asian countries. Although Japan is known to achieve high survival from stomach cancer, little is known regarding the survival trends for recent years and survival by subsite and stage. We report age-standardised 1-, 3-, 5- and 10-year net survival for patients diagnosed with stomach cancer in Osaka, Japan.MethodsWe analysed patients diagnosed with primary stomach cancer and registered in the population-based cancer registry in Osaka Prefecture between 2001 and 2014. We used the non-parametric Pohar Perme method to derive net survival for each year. Both cohort and period approaches were used. Age was standardised using weights of the external population of the International Cancer Survival Standard. Multiple imputation was applied to handle missing information on subsite and stage before estimating age-standardised net survival by subsite (cardia and non-cardia) and stage (localised, regional and distant metastasis). We then examined general trends in the cohort-based survival estimates, as well as by subsite and stage, using linear regression.ResultsA total of 97,276 patients were included in the analysis. Age-standardised net survival improved steadily (mean annual absolute change ≥1.2%). Net survival for both subsites improved, but cardia cancer showed 7–23% lower survival than non-cardia cancer throughout the study period. Five-year net survival remained high (≥80%) in the localised stage from the beginning of this study. Net survival increased steeply (≥1.4% per year) in the regional stage. Although 1-year net survival increased by 14% in the distant stage, 5-year and 10-year net survival remained below 10%.ConclusionAge-standardised net survival for stomach cancer in Japan improved during the study period owing to an increase in the number of patients with localised stage at diagnosis and improved treatment. Monitoring both short- and long-term survival should be continued as management of stomach cancer progresses.     

Surgical resection as a predictor of cancer-specific survival by stage at diagnosis and cancer type,United States, 2006–2015

BackgroundWhen solid tumors are amenable to definitive resection, clinical outcomes are generally superior to when those tumors are inoperable. However, the population-level cancer survival benefit of eligibility for surgery by cancer stage has not yet been quantified.MethodsUsing Surveillance, Epidemiology and End Results data allowing us to identify patients who were deemed eligible for and received surgical resection, we examined the stage-specific association of surgical resection with 12-year cancer-specific survival. The 12-year endpoint was selected to maximize follow-up time and thereby minimize the influence of lead time bias.ResultsAcross a variety of solid tumor types, earlier stage at diagnosis allowed for surgical intervention at a much higher rate than later-stage diagnosis. At every stage, surgical intervention was associated with a substantially higher rate of 12-year cancer-specific survival, with absolute differences of up to 51% for stage I, 51% for stage II, and 44% for stage III cancer, and stage-specific mortality relative risks of 3.6, 2.4, and 1.7, respectively.ConclusionsDiagnosis of solid cancers in early stages often enables surgical resection, which reduces the risk of death from cancer. Receipt of surgical resection is an informative endpoint that is strongly associated with long-term cancer-specific survival at every stage.     

Simple parametric survival analysis with anonymized register data: A cohort study with truncated and interval censored event and censoring times

Background

To preserve patient anonymity, health register data may be provided as binned data only. Here we consider as example, how to estimate mean survival time after a diagnosis of metastatic colorectal cancer from Norwegian register data on time to death or censoring binned into 30 day intervals. All events occurring in the first three months (90 days) after diagnosis were removed to achieve comparability with a clinical trial. The aim of the paper is to develop and implement a simple, and yet flexible method for analyzing such interval censored and truncated data.

Methods

Considering interval censoring a missing data problem, we implement a simple multiple imputation strategy that allows flexible sensitivity analyses with respect to the shape of the censoring distribution. To allow identification of appropriate parametric models, a χ²-goodness-of-fit test--also imputation based--is derived and supplemented with diagnostic plots. Uncertainty estimates for mean survival times are obtained via a simulation strategy. The validity and statistical efficiency of the proposed method for varying interval lengths is investigated in a simulation study and compared with simpler alternatives.

Results

Mean survival times estimated from the register data ranged from 1.2 (SE = 0.09) to 3.2 (0.31) years depending on period of diagnosis and choice of parametric model. The shape of the censoring distribution within intervals did generally not influence results, whereas the choice of parametric model did, even when different models fit the data equally well. In simulation studies both simple midpoint imputation and multiple imputation yielded nearly unbiased analyses (relative biases of -0.6% to 9.4%) and confidence intervals with near-nominal coverage probabilities (93.4% to 95.7%) for censoring intervals shorter than six months. For 12 month censoring intervals, multiple imputation provided better protection against bias, and coverage probabilities closer to nominal values than simple midpoint imputation.

Conclusion

Binning of event and censoring times should be considered a viable strategy for anonymizing register data on survival times, as they may be readily analyzed with methods based on multiple imputation.

     

Performance of the marginal structural models under various scenarios of incomplete marker's values: A simulation study

Marginal structural models (MSMs) have been proposed for estimating a treatment's effect, in the presence of time‐dependent confounding. We aimed to evaluate the performance of the Cox MSM in the presence of missing data and to explore methods to adjust for missingness. We simulated data with a continuous time‐dependent confounder and a binary treatment. We explored two classes of missing data: (i) missed visits, which resemble clinical cohort studies; (ii) missing confounder's values, which correspond to interval cohort studies. Missing data were generated under various mechanisms. In the first class, the source of the bias was the extreme treatment weights. Truncation or normalization improved estimation. Therefore, particular attention must be paid to the distribution of weights, and truncation or normalization should be applied if extreme weights are noticed. In the second case, bias was due to the misspecification of the treatment model. Last observation carried forward (LOCF), multiple imputation (MI), and inverse probability of missingness weighting (IPMW) were used to correct for the missingness. We found that alternatives, especially the IPMW method, perform better than the classic LOCF method. Nevertheless, in situations with high marker's variance and rarely recorded measurements none of the examined method adequately corrected the bias.     

Subsite- and stage-specific colorectal cancer trends in Estonia prior to implementation of screening

BackgroundThe occurrence of colorectal cancer (CRC) in Estonia has been characterised by increasing incidence, low survival and no screening. The study aimed to examine long-term incidence and survival trends of CRC in Estonia with specific focus on subsite and stage.MethodsWe analysed CRC incidence and relative survival using Estonian Cancer Registry data on all cases of colorectal cancer (ICD-10 C18–21) diagnosed in 1995–2014. TNM classification was used to categorise stage.ResultsAge-standardized incidence of colon cancer increased both in men and women at a rate of approximately 1% per year. Significant increase was seen for right-sided tumours, but not for left-sided tumours. Rectal cancer incidence increased significantly only in men and anal cancer incidence only in women. Age-standardized five-year relative survival for colon cancer increased from 50% in 1995–1999 to 59% in 2010–2014; for rectal cancer, from 38% to 56%. Colon cancer survival improved significantly for left-sided tumours (from 51% to 62%) and stage IV disease (from 6% to 15%). For rectal cancer, significant survival gain was seen for stage II (from 58% to 75%), stage III (from 34% to 70%) and stage IV (from 1% to 12%).ConclusionIn the pre-screening era in Estonia, increase in colon cancer incidence was limited to right-sided tumours. Large stage-specific survival gain, particularly for rectal cancer, was probably due to better staging and advances in multimodality treatment. Nonetheless, more than one quarter of new CRC cases are diagnosed at stage IV, emphasising the need for an efficient screening program.     

If a picture is worth a thousand words,take a good look at the picture: Survival after liver metastasectomy for colorectal cancer

IntroductionAn analysis of NHS data published in by Morris et al. in 2010 is widely used as evidence in support of liver metastasectomy for colorectal cancer and its wider application. Recent evidence concerning better overall survival for patients with metastatic colorectal cancer challenges the notional assumptions about what survival would be without metastasectomy. Earlier detection of metastases for local treatments has not resulted in a survival benefit.Materials and methodsThe interpretation of its central graphical display is critically reviewed and the common the limitations of the analysis of registry data and resulting immortal time bias are explored.Results and discussionRecent evidence, including the 2017 CLOCC trial report make the original interpretation of the analysis suspect. Randomised trials are essential to detect a treatment effect of specific interventions among variable disease progression, selection bias, and multiple and repeated treatments that are inherent in the management of advanced cancer.     

Model selection of generalized estimating equations with multiply imputed longitudinal data

Longitudinal data often encounter missingness with monotone and/or intermittent missing patterns. Multiple imputation (MI) has been popularly employed for analysis of missing longitudinal data. In particular, the MI‐GEE method has been proposed for inference of generalized estimating equations (GEE) when missing data are imputed via MI. However, little is known about how to perform model selection with multiply imputed longitudinal data. In this work, we extend the existing GEE model selection criteria, including the “quasi‐likelihood under the independence model criterion” (QIC) and the “missing longitudinal information criterion” (MLIC), to accommodate multiple imputed datasets for selection of the MI‐GEE mean model. According to real data analyses from a schizophrenia study and an AIDS study, as well as simulations under nonmonotone missingness with moderate proportion of missing observations, we conclude that: (i) more than a few imputed datasets are required for stable and reliable model selection in MI‐GEE analysis; (ii) the MI‐based GEE model selection methods with a suitable number of imputations generally perform well, while the naive application of existing model selection methods by simply ignoring missing observations may lead to very poor performance; (iii) the model selection criteria based on improper (frequentist) multiple imputation generally performs better than their analogies based on proper (Bayesian) multiple imputation.     

Cure by age and stage at diagnosis for colorectal cancer patients in North West England, 1997–2004: A population-based study

Background: Stage and age at diagnosis are important prognostic factors for patients with colorectal cancer. However, the proportion cured by stage and age is unknown in England. Materials and methods: This population-based study includes 29,563 adult patients who were diagnosed and registered with colorectal cancer during 1997–2004 and followed till 2007 in North West England. Multiple imputation was used to provide more reliable estimates of stage at diagnosis, when these data were missing. Cure mixture models were used to estimate the proportion ‘cured’ and the median survival of the uncured by age and stage. Results: For both colon and rectal cancer the proportion of patients cured and median survival time of the uncured decreased with advancing stage and increasing age. Patients aged under 65 years had the highest proportion cured and longest median survival of the uncured. Conclusion: Cure of colorectal cancer patients is dependent on stage and age at diagnosis with younger patients or those with less advanced disease having a better prognosis. Further efforts are required, in order to reduce the proportion of patients presenting with stage III and IV disease and ultimately increase the chance of cure.     

Viral Genetic Linkage Analysis in the Presence of Missing Data

Analyses of viral genetic linkage can provide insight into HIV transmission dynamics and the impact of prevention interventions. For example, such analyses have the potential to determine whether recently-infected individuals have acquired viruses circulating within or outside a given community. In addition, they have the potential to identify characteristics of chronically infected individuals that make their viruses likely to cluster with others circulating within a community. Such clustering can be related to the potential of such individuals to contribute to the spread of the virus, either directly through transmission to their partners or indirectly through further spread of HIV from those partners. Assessment of the extent to which individual (incident or prevalent) viruses are clustered within a community will be biased if only a subset of subjects are observed, especially if that subset is not representative of the entire HIV infected population. To address this concern, we develop a multiple imputation framework in which missing sequences are imputed based on a model for the diversification of viral genomes. The imputation method decreases the bias in clustering that arises from informative missingness. Data from a household survey conducted in a village in Botswana are used to illustrate these methods. We demonstrate that the multiple imputation approach reduces bias in the overall proportion of clustering due to the presence of missing observations.     

Socioeconomic status and cancer survival in Brazil: Analysis of population data from the municipalities of Aracaju and Curitiba, 1996–2012

IntroductionThe association between socioeconomic status and cancer prognosis has been demonstrated in several countries. Despite the existence of indirect evidence of this phenomenon in Brazil, few studies in this regard are available.ObjectivesThe objective of the present study is to analyse socioeconomic related survival gaps for patients diagnosed with breast, cervical, lung, prostate, and colorectal cancer in the cities of Aracaju (SE) and Curitiba (PR).MethodsUsing population-based data, we estimated net survival by tumour site, year of diagnosis, socioeconomic status and local of residence. Net survival estimation was done with multilevel parametric model allowing flexible spline functions do estimate excess mortality hazards.Results28,005 cases were included in survival analysis. Five-year net survival showed positive association with SES. Intermunicipal survival gaps favouring Aracaju where prominent for breast (reaching 16,1% in 5 years)ObjectivesStudy the impact of socioeconomic factors on cancer survival in two Brazilian capitals. Methods: Survival analysis using population-based cancer data including patients diagnosed with breast, lung, prostate, cervical and colorectal cancer between 1996 and 2012 in Aracaju and Curitiba. Outcomes were excessive mortality hazard (EMH) and 5- and 8-years net survival (NS). The association of race/skin color and socioeconomic level (SES) with EMH and net survival were analyzed using a multilevel regression model with flexible splines.Results28,005 cases were included, 6636 from Aracaju and 21,369 from Curitiba. NS for all diseases studied increased more prominently for Curitiba population. We observed NS gap between the populations of Aracaju and Curitiba that increased or remained stable during the study period, with emphasis on the growth of the difference in NS of lung and colon cancer (among men). Only for cervical cancer and prostate cancer there was a reduction in the intermunicipal gaps. 5-year NS for breast cancer in Aracaju ranged from 55.2% to 73.4% according to SES. In Curitiba this variation was from 66.5% to 83.8%.ConclusionThe results of the present study suggests widening of socioeconomic and regional inequalities in the survival of patients with colorectal, breast, cervical, lung and prostate cancers in Brazil during the 1990 s and 2000 s     

Statistical Inference for Hardy-Weinberg Proportions in the Presence of Missing Genotype Information

In genetic association studies, tests for Hardy-Weinberg proportions are often employed as a quality control checking procedure. Missing genotypes are typically discarded prior to testing. In this paper we show that inference for Hardy-Weinberg proportions can be biased when missing values are discarded. We propose to use multiple imputation of missing values in order to improve inference for Hardy-Weinberg proportions. For imputation we employ a multinomial logit model that uses information from allele intensities and/or neighbouring markers. Analysis of an empirical data set of single nucleotide polymorphisms possibly related to colon cancer reveals that missing genotypes are not missing completely at random. Deviation from Hardy-Weinberg proportions is mostly due to a lack of heterozygotes. Inbreeding coefficients estimated by multiple imputation of the missings are typically lowered with respect to inbreeding coefficients estimated by discarding the missings. Accounting for missings by multiple imputation qualitatively changed the results of 10 to 17% of the statistical tests performed. Estimates of inbreeding coefficients obtained by multiple imputation showed high correlation with estimates obtained by single imputation using an external reference panel. Our conclusion is that imputation of missing data leads to improved statistical inference for Hardy-Weinberg proportions.     

Changes in prostate cancer incidence,mortality and survival in relation to prostate specific antigen testing in New South Wales,Australia

BackgroundTo examine changes in prostate cancer incidence and mortality rates, and 5-year relative survival, in relation to changes in the rate of prostate specific antigen (PSA) screening tests and the use of radical prostatectomy (RP) in the Australian population.MethodsProstate cancer stage-specific incidence rates, 5-year relative survival and mortality rates were estimated using New South Wales Cancer Registry data. PSA screening test rates and RP/Incidence ratios were estimated from Medicare Benefits Schedule claims data. We used multiple imputation to impute stage for cases with “unknown” stage at diagnosis. Annual percentage changes (APC) in rates were estimated using Joinpoint regression.ResultsTrends in the age-standardized incidence rates for localized disease largely mirrored the trends in PSA screening test rates, with a substantial ‘spike’ in the rates occurring in 1994, followed by a second ‘spike’ in 2008, and then a significant decrease from 2008 to 2015 (APC −6.7, 95% CI −8.2, −5.1). Increasing trends in incidence rates were observed for regional stage from the early 2000s, while decreasing or stable trends were observed for distant stage since 1993. The overall RP/Incidence ratio increased from 1998 to 2003 (APC 9.6, 95% CI 3.8, 15.6), then remained relatively stable to 2015. The overall 5-year relative survival for prostate cancer increased from 58.4% (95% CI: 55.0–61.7%) in 1981–1985 to 91.3% (95% CI: 90.5–92.1%) in 2011–2015. Prostate cancer mortality rates decreased from 1990 onwards (1990–2006: APC −1.7, 95% CI −2.1, −1.2; 2006–2017: APC −3.8, 95% CI −4.4, −3.1).ConclusionsOverall, there was a decrease in the incidence rate of localized prostate cancer after 2008, an increase in survival over time and a decrease in the mortality rate since the 1990s. This seems to indicate that the more conservative use of PSA screening tests in clinical practice since 2008 has not had a negative impact on population-wide prostate cancer outcomes.     

Colorectal cancer and country of birth in New South Wales,Australia: All-of-population data for prioritising health service delivery and research

IntroductionCancer care and outcomes differ across cultural groups in Australia. Quantifying these differences facilitates prioritisation and targeting of services and research. All-of-population data are needed by health agencies to understand and fulfil their cancer-control responsibilities. Compiling these data can be challenging while maintaining privacy. We have used data linkage to gain population-wide colorectal cancer data on stage (degree of spread), treatment, and survival in New South Wales (NSW), Australia, by country of birth (COB), and consider service implications.MethodsWe studied colon and rectal cancers diagnosed in 2003–2016 and recorded on the NSW Cancer Registry (n = 41,575), plus linked hospital data and data from Australian Medical and Pharmaceutical Benefits payments, other treatment data and death records. Outcomes for 12 COB categories were analysed using multiple logistic and proportional hazards regression, with Australia as the reference category.ResultsCompared with Australian born, the adjusted odds ratio for distant spread of colon cancer was higher for people born in Lebanon and the United Kingdom. Treatment was less common for people born in China (surgery), Germany (systemic), Italy (surgery), New Zealand (any treatment) and Vietnam (all treatments), while treatment for rectal cancer was more common for people born in Italy (surgery), United Kingdom (radiotherapy, systemic therapy), and Vietnam (surgery), and less frequent for people born in China (radiotherapy). Adjusted 5-year survival was higher for people born in China, Italy, Vietnam, Greece (colon), Lebanon (colon) and other non-English speaking countries. More advanced stage was negatively related to having surgery and survival.ConclusionsThis study illustrates how linked data can enable comparisons of multiple outcomes for colorectal cancer by country of birth across an entire population. Results disclose “big picture” variations in population characteristics, stage, treatment and survival. This will enable better targeting and prioritisation of services and inform research priorities to address disparities.     

A Semiparametric Missing‐Data‐Induced Intensity Method for Missing Covariate Data in Individually Matched Case–Control Studies

Summary In individually matched case–control studies, when some covariates are incomplete, an analysis based on the complete data may result in a large loss of information both in the missing and completely observed variables. This usually results in a bias and loss of efficiency. In this article, we propose a new method for handling the problem of missing covariate data based on a missing‐data‐induced intensity approach when the missingness mechanism does not depend on case–control status and show that this leads to a generalization of the missing indicator method. We derive the asymptotic properties of the estimates from the proposed method and, using an extensive simulation study, assess the finite sample performance in terms of bias, efficiency, and 95% confidence coverage under several missing data scenarios. We also make comparisons with complete‐case analysis (CCA) and some missing data methods that have been proposed previously. Our results indicate that, under the assumption of predictable missingness, the suggested method provides valid estimation of parameters, is more efficient than CCA, and is competitive with other, more complex methods of analysis. A case–control study of multiple myeloma risk and a polymorphism in the receptor Inter‐Leukin‐6 (IL‐6‐α) is used to illustrate our findings.