This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Electronic health records are a valuable data source for investigating health-related questions, and propensity score analysis has become an increasingly popular approach to address confounding bias in such investigations. However, because electronic health records are typically routinely recorded as part of standard clinical care, there are often missing values, particularly for potential confounders. In our motivating study-using electronic health records to investigate the effect of renin-angiotensin system blockers on the risk of acute kidney injury-two key confounders, ethnicity and chronic kidney disease stage, have 59% and 53% missing data, respectively. The missingness pattern approach (MPA), a variant of the missing indicator approach, has been proposed as a method for handling partially observed confounders in propensity score analysis. In the MPA, propensity scores are estimated separately for each missingness pattern present in the data. Although the assumptions underlying the validity of the MPA are stated in the literature, it can be difficult in practice to assess their plausibility. In this article, we explore the MPA's underlying assumptions by using causal diagrams to assess their plausibility in a range of simple scenarios, drawing general conclusions about situations in which they are likely to be violated. We present a framework providing practical guidance for assessing whether the MPA's assumptions are plausible in a particular setting and thus deciding when the MPA is appropriate. We apply our framework to our motivating study, showing that the MPA's underlying assumptions appear reasonable, and we demonstrate the application of MPA to this study.

Objective: To compare the management and outcomes of colorectal cancer (CRC) patients during the first 2 months of the COVID-19 pandemic with the preceding 6 months. Background: The pandemic has affected the diagnosis and treatment of CRC patients worldwide. Little is known about the safety of major resection and whether creating “cold” sites (COVID-free hospitals) is effective. Methods: A national study in England used administrative hospital data for 14,930 CRC patients undergoing surgery between October 1, 2019, and May 31, 2020. Mortality of CRC resection was compared before and after March 23, 2020 (“lockdown” start). Results: The number of elective CRC procedures dropped sharply during the pandemic (from average 386 to 214 per week), whereas emergency procedures were hardly affected (from 88 to 84 per week). There was little change in characteristics of surgical patients during the pandemic. Laparoscopic surgery decreased from 62.5% to 35.9% for elective and from 17.7% to 9.7% for emergency resections. Surgical mortality increased slightly (from 0.9% to 1.2%, P = 0.06) after elective and markedly (from 5.6% to 8.9%, P = 0.003) after emergency resections. The observed increase in mortality during the first phase of the pandemic was similar in “cold” and “hot” sites ( P > 0.5 elective and emergency procedures). Conclusions: The pandemic resulted in a 50% reduction in elective CRC procedures during the initial surge and a substantial increase in mortality after emergency resection. There was no evidence that surgery in COVID-free “cold” sites led to better outcomes in the first 2 months.

Background Multivariate meta‐analysis (MVMA) jointly synthesizes effects for multiple correlated outcomes. The MVMA model is potentially more difficult and time‐consuming to apply than univariate models, so if its use makes little difference to parameter estimates, it could be argued that it is redundant. Methods We assessed the applicability and impact of MVMA in Cochrane Pregnancy and Childbirth (CPCB) systematic reviews. We applied MVMA to CPCB reviews published between 2011 and 2013 with two or more binary outcomes with at least three studies and compared findings with results of univariate meta‐analyses. Univariate random effects meta‐analysis models were fitted using restricted maximum likelihood estimation (REML). Results Eighty CPCB reviews were published. MVMA could not be applied in 70 of these reviews. MVMA was not feasible in three of the remaining 10 reviews because the appropriate models failed to converge. Estimates from MVMA agreed with those of univariate analyses in most of the other seven reviews. Statistical significance changed in two reviews: In one, this was due to a very small change in P value; in the other, the MVMA result for one outcome suggested that previous univariate results may be vulnerable to small‐study effects and that the certainty of clinical conclusions needs consideration. Conclusions MVMA methods can be applied only in a minority of reviews of interventions in pregnancy and childbirth and can be difficult to apply because of missing correlations or lack of convergence. Nevertheless, clinical and/or statistical conclusions from MVMA may occasionally differ from those from univariate analyses.

Missing data is a common issue in research using observational studies to investigate the effect of treatments on health outcomes. When missingness occurs only in the covariates, a simple approach is to use missing indicators to handle the partially observed covariates. The missing indicator approach has been criticized for giving biased results in outcome regression. However, recent papers have suggested that the missing indicator approach can provide unbiased results in propensity score analysis under certain assumptions. We consider assumptions under which the missing indicator approach can provide valid inferences, namely, (1) no unmeasured confounding within missingness patterns; either (2a) covariate values of patients with missing data were conditionally independent of treatment or (2b) these values were conditionally independent of outcome; and (3) the outcome model is correctly specified: specifically, the true outcome model does not include interactions between missing indicators and fully observed covariates. We prove that, under the assumptions above, the missing indicator approach with outcome regression can provide unbiased estimates of the average treatment effect. We use a simulation study to investigate the extent of bias in estimates of the treatment effect when the assumptions are violated and we illustrate our findings using data from electronic health records. In conclusion, the missing indicator approach can provide valid inferences for outcome regression, but the plausibility of its assumptions must first be considered carefully.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Aim:The aim was to compare early postoperative outcomes and 2-year cancer-specific mortality following colorectal cancer (CRC) resection in patients with and without inflammatory bowel disease (IBD) in England and Wales.Method: Records for patients in the National Bowel Cancer Audit who had major CRC resection between April 2014 and December 2017 were linked to routinely collected hospital level administrative datasets and chemotherapy and radiotherapy datasets.Multivariable regression models were used to compare outcomes with adjustment for patient and tumour characteristics.Results: In all, 63 365 patients were included. 1285 (2.0%) had an IBD diagnosis: 839 (65.3%) ulcerative colitis, 435 (33.9%) Crohn's disease and 11 (0.9%) were indeterminate.IBD patients were younger, had more advanced cancer staging and a higher proportion of right-sided tumours. They also had a higher proportion of emergency resection, total/ subtotal colectomy, open surgery and stoma formation at resection, with longer hospital admissions and higher rates of unplanned readmission and reoperation. Fewer rectal cancer patients with IBD received neoadjuvant radiotherapy (24.8% vs. 36.0%, P = 0.005) whilst similar proportions of Stage III colon cancer patients received adjuvant chemotherapy. Ninety-day postoperative mortality was similar, but unadjusted 2-year cancerspecific mortality was significantly higher in patients with IBD (subdistribution hazard ratio 1.35, 95% CI 1.18-1.55). Risk adjustment for patient and tumour factors reduced this association (adjusted subdistribution hazard ratio 1.22, 95% CI 1.05-1.43). Conclusion:Patients with IBD and CRC are a distinct patient group who develop CRC at a younger age and undergo more radical surgery. They have worse cancer survival, with the difference in prognosis appearing after the early postoperative period.

Background Methods for linking records between two datasets are well established. However, guidance is needed for linking more than two datasets. Using all ‘pairwise linkages’—linking each dataset to every other dataset—is the most inclusive, but resource-intensive, approach. The ‘spine’ approach links each dataset to a designated ‘spine dataset’, reducing the number of linkages, but potentially reducing linkage quality. Methods We compared the pairwise and spine linkage approaches using real-world data on patients undergoing emergency bowel cancer surgery between 31 October 2013 and 30 April 2018. We linked an administrative hospital dataset (Hospital Episode Statistics; HES) capturing patients admitted to hospitals in England, and two clinical datasets comprising patients diagnosed with bowel cancer and patients undergoing emergency bowel surgery. Results The spine linkage approach, with HES as the spine dataset, created an analysis cohort of 15 826 patients, equating to 98.3% of the 16 100 patients identified using the pairwise linkage approach. There were no systematic differences in patient characteristics between these analysis cohorts. Associations of patient and tumour characteristics with mortality, complications and length of stay were not sensitive to the linkage approach. When eligibility criteria were applied before linkage, spine linkage included 14 509 patients (90.0% compared with pairwise linkage). Conclusion Spine linkage can be used as an efficient alternative to pairwise linkage if case ascertainment in the spine dataset and data quality of linkage variables are high. These aspects should be systematically evaluated in the nominated spine dataset before spine linkage is used to create the analysis cohort.