BACKGROUND: Empirical evaluation of inverse probability weighting (IPW) for self-selection bias correction is inaccessible without the full source population. We aimed to: (i) investigate how self-selection biases frequency and association measures and (ii) assess self-selection bias correction using IPW in a cohort with register linkage.
METHODS: The source population included 17 936 individuals invited to the Copenhagen Aging and Midlife Biobank during 2009-11 (ages 49-63 years). Participants counted 7185 (40.1%). Register data were obtained for every invited person from 7 years before invitation to the end of 2020. The association between education and mortality was estimated using Cox regression models among participants, IPW participants and the source population.
RESULTS: Participants had higher socioeconomic position and fewer hospital contacts before baseline than the source population. Frequency measures of participants approached those of the source population after IPW. Compared with primary/lower secondary education, upper secondary, short tertiary, bachelor and master/doctoral were associated with reduced risk of death among participants (adjusted hazard ratio [95% CI]: 0.60 [0.46; 0.77], 0.68 [0.42; 1.11], 0.37 [0.25; 0.54], 0.28 [0.18; 0.46], respectively). IPW changed the estimates marginally (0.59 [0.45; 0.77], 0.57 [0.34; 0.93], 0.34 [0.23; 0.50], 0.24 [0.15; 0.39]) but not only towards those of the source population (0.57 [0.51; 0.64], 0.43 [0.32; 0.60], 0.38 [0.32; 0.47], 0.22 [0.16; 0.29]).
CONCLUSIONS: Frequency measures of study participants may not reflect the source population in the presence of self-selection, but the impact on association measures can be limited. IPW may be useful for (self-)selection bias correction, but the returned results can still reflect residual or other biases and random errors.

BACKGROUND: Mental health disorders are common among people in prison, but their prevalence in the Scandinavian prison population remain unclear. In this multinational register study, we examined the prevalence of mental health disorders and the comorbidity of substance use disorders (SUDs) with other mental health disorders in this population. Further, we investigated how the prevalence of mental disorders at prison entry had changed in Norway, Denmark, and Sweden over the study period.
METHODS: The three study cohorts included all individuals, aged 19 or older, whom had been imprisoned in Norway (2010-2019), Denmark (2011-2018), and Sweden (2010-2013). Mental disorders were defined as ICD-10 diagnoses (F-codes) registered in the national patient registers. The study prevalence was estimated based on recorded diagnoses during the entire study follow-up period in each respective country. The one-year prevalence of mental disorders was estimated for each calendar year for individuals entering prison during that year.
RESULTS: The Scandinavian prison cohorts included 119 507 individuals released 191 549 times during the study period. Across all three countries a high proportion of both women (61.3%-74.4%) and men (49.6%-57.9%) had at least one mental health disorder during the observation period. The most prevalent disorders were SUDs (39.1%-44.0%), depressive disorder (8.1%-17.5%), and stress related disorder (8.8%-17.1%). Women (31.8%-41.1%) had higher levels of mental health and substance use comorbidities compared to men (20.8%-27.6%). The one-year prevalence of any mental health disorder increased over time with a 33% relative increase in Norway, 8% in Denmark, and 10% in Sweden. The proportion of individuals entering prison with a comorbid SUD and other mental disorder had also increased.
CONCLUSIONS: While the incarceration rate has been decreasing during the past decade in the Scandinavian countries, an increasing proportion of people entering prison have a diagnosed mental health disorder. Our results suggest that prisons should provide adequate treatment and scale up services to accommodate the increasing proportion of people with complex health needs among incarcerated people.

BACKGROUND: Deterioration of glycaemic control in people with long-standing diabetes mellitus (diabetes) may be a possible indicator of pancreatic cancer. However, the magnitude of the association between diabetes deterioration and pancreatic cancer has received little attention.
METHODS: We conducted a matched cohort study, nested within a population-based cohort of Australian women with diabetes. Women with unstable diabetes, defined as a change in medication after a 2-year period of stable medication use, were matched by birth year to those with stable diabetes, in a 1:4 ratio. We used flexible parametric survival models to estimate hazard ratios (HRs) and 95% confidence intervals (CI).
RESULTS: We included 134,954 and 539,789 women in the unstable and stable diabetes cohorts, respectively (mean age 68 years). In total, 1,315 pancreatic cancers were diagnosed. Deterioration of stable diabetes was associated with a 2.5-fold increased risk of pancreatic cancer (HR 2.55; 95% CI 2.29-2.85). The risk was particularly high within the first year after diabetes deteriorated. HRs at 3 months, 6 months and 1 year were: 5.76 (95% CI 4.72-7.04); 4.56 (95% CI 3.81-5.46); and 3.33 (95% CI 2.86-3.89), respectively. The risk was no longer significantly different after 7 years.
CONCLUSIONS: Deterioration in glycaemic control in people with previously stable diabetes may be an indicator of pancreatic cancer, suggesting investigations of the pancreas may be appropriate. The weaker longer-term (3-7 years) association between diabetes deterioration and pancreatic cancer may indicate that poor glycaemic control can be a risk factor for pancreatic cancer.

OBJECTIVE: The bidirectional association between diabetes mellitus (DM) and pancreatic cancer (PC) is established; however, the strength of association between duration of DM and risk of PC needs further investigation.
METHODS: We conducted a case-control study nested within a population-based cohort of Australian women established using record linkage. Women diagnosed with PC from July 2007 to December 2013, were matched to five controls based on age and state of residence. DM was defined according to prescription of anti-diabetic medication from administrative prescription data. We used conditional logistic regression to calculate odds ratios (OR) and 95% confidence intervals (CI), adjusted for area-level socioeconomic status, rurality of residence, weighted comorbidity score, and predicted probability of obesity.
RESULTS: The analyses included 7,267 cases and 35,978 controls. The mean age at the time of DM diagnosis was 71 years whereas the mean age at the time of diagnosis of PC was 76 years. A history of DM of any duration was associated with a 2-fold increase in risk of PC (OR=2.12; 95%CI:1.96-2.29) compared to having no history of DM. The risk decreased with increasing duration of DM. The highest risk was in those who had recent-onset DM (OR=8.08; 95%CI:6.88-9.50 for <12 months of DM), but the risk remained elevated with ≥5 years of DM (OR=1.40; 95%CI:1.27-1.55).
CONCLUSIONS: The markedly increased risk of PC in those with recent-onset DM emphasises the need for further research to distinguish patients for whom new-onset DM is a manifestation of PC from those with type-2 DM. The elevated risk associated with long-standing DM suggests that preventing DM may contribute to a reduction in the incidence of PC.

OBJECTIVE: A recent paper suggested all women with endometrial cancer should take statins but it is unclear whether there is sufficient evidence to justify this recommendation.
METHODS: We identified all women diagnosed with uterine cancer in Australia between July 2003 and December 2013 (2012 in New South Wales) through the Australian Cancer Database (N = 16,501) and linked these to the national prescription database and National Death Index to identify statin use and survival outcomes to December 2015. We used Cox proportional hazards regression to estimate hazard ratios (HR) and 95% confidence intervals (CI) for the associations between statin use and survival.
RESULTS: Among the 15,703 women with endometrial cancer, pre-diagnosis statin use was not associated with survival. Endometrial cancer-specific mortality was lower among women who used statins after diagnosis (time-varying models: HR = 0.92; 95%CI 0.82-1.03) but the association was only seen among women with type 1 cancers (0.87; 0.76-1.00), for hydrophilic statins (0.84; 0.68-1.03) and for new use of statins after diagnosis (0.75; 0.59-0.95). There was a weak dose-response with increasing number of statin prescriptions. Sensitivity analyses using inverse probability of treatment weights were similar.
CONCLUSIONS: Women with endometrial cancer who take statins after diagnosis may have better survival than those who do not use statins. However, it is impossible to completely rule out bias, particularly reverse causation where disease status may affect statin use. We believe it is too early to recommend all women with endometrial cancer take statins, but there is sufficient evidence to justify a randomized trial.

We introduce and study a recently proposed method for privacy-preserving distance computations which has received little attention in the scientific literature so far. The method, which is based on intersecting sets of randomly labeled grid points, is henceforth denoted as ISGP allows calculating the approximate distances between masked spatial data. Coordinates are replaced by sets of hash values. The method allows the computation of distances between locations L when the locations at different points in time t are not known simultaneously. The distance between [Formula: see text] and [Formula: see text] could be computed even when [Formula: see text] does not exist at [Formula: see text] and [Formula: see text] has been deleted at [Formula: see text]. An example would be patients from a medical data set and locations of later hospitalizations. ISGP is a new tool for privacy-preserving data handling of geo-referenced data sets in general. Furthermore, this technique can be used to include geographical identifiers as additional information for privacy-preserving record-linkage. To show that the technique can be implemented in most high-level programming languages with a few lines of code, a complete implementation within the statistical programming language R is given. The properties of the method are explored using simulations based on large-scale real-world data of hospitals ([Formula: see text]) and residential locations ([Formula: see text]). The method has already been used in a real-world application.
ISGP yields very accurate results. Our simulation study showed that-with appropriately chosen parameters - 99 % accuracy in the approximated distances is achieved.
We discussed a new method for privacy-preserving distance computations in microdata. The method is highly accurate, fast, has low computational burden, and does not require excessive storage.

BACKGROUND: Despite the steady decline in the last few decades, Portugal remains the Western European country with the highest TB notification rates. The aim of this study was to estimate the completeness of notification to the National Tuberculosis Programme (NTP) Surveillance System (SVIG-TB) in 2015.METHODS: We implemented an inventory study and a three-source log-linear capture-recapture analysis using two additional data sources that were deterministic and probabilistically linked: the national notifiable diseases surveillance system (Sistema Nacional de Vigilância Epidemiológica SINAVE) and the national hospital discharge database (Grupos de Diagnósticos Homogéneos GDH).RESULTS: We identified 2328 unique probable/confirmed TB cases across the three data sources. We found a positive dependency between SVIG-TB and SINAVE (incidence rate ratio IRR 8.9, 95%CI 6.6-12.0) and between GDH and SINAVE (IRR 2.6, 95%CI 2.0-3.4). After adjusting for these dependencies, we estimated that 266 cases (95%CI 198-358) were not reported, indicating a notification (to SVIG-TB) completeness rate of 77.0%.CONCLUSION: True incidence rate of TB in Portugal in 2015 could have been as high as 26.1 per 100 000. This could be an overestimation because of false-positive cases recorded in both SINAVE and GDH or on a smaller scale, false non-matches. Studies aimed at validating potentially false-positive cases should be implemented to address these limitations.

Insulin-like growth factor-I (IGF-I) and testosterone may be related to prostate cancer risk. Acromegaly is associated with clinically high IGF-I concentrations. Klinefelter\'s syndrome, testicular hypofunction and hypopituitarism are associated with clinically low testosterone concentrations. We aimed to investigate whether diagnosis with these conditions was associated with subsequent prostate cancer diagnosis and mortality. We used linked English national Hospital Episode Statistics and mortality data from 1999 to 2017 to construct and follow-up cohorts of men aged ≥35 years diagnosed with (i) acromegaly (n = 2,495) and (ii) hypogonadal-associated diseases (n = 18,763): Klinefelter\'s syndrome (n = 1,992), testicular hypofunction (n = 8,086) and hypopituitarism (n = 10,331). We estimated adjusted hazard ratios (HRs) and confidence intervals (CIs) for prostate cancer diagnosis and death using Cox regression in comparison with an unexposed reference cohort of 4.3 million men, who were admitted to hospital for a range of minor surgeries and conditions (n observed cases = 130,000, n prostate cancer deaths = 30,000). For men diagnosed with acromegaly, HR for prostate cancer diagnosis was 1.33 (95% CI 1.09-1.63; p = 0.005; n observed cases = 96), HR for prostate cancer death was 1.44 (95% CI 0.92-2.26; p = 0.11; n deaths = 19). Diagnosis with Klinefelter\'s syndrome was associated with a lower prostate cancer risk (HR = 0.58, 95% CI 0.37-0.91; p = 0.02; n observed cases = 19) and hypopituitarism was associated with a reduction in prostate cancer death (HR = 0.53, 95% CI 0.35-0.79; p = 0.002; n deaths = 23). These results support the hypothesised roles of IGF-I and testosterone in prostate cancer development and/or progression. These findings are important because they provide insight into prostate cancer aetiology.

Rotavirus vaccines (RV), included in Australia\'s National Immunisation Program from mid-July 2007, are unique in strict time limits for administration. Here, we report on timeliness of RV uptake, compare cumulative RV coverage to age 12 months with DTPa, and assess factors associated with receipt of RV among Aboriginal and non-Aboriginal children.
Birth records for 681,456 children born in two Australian states in 2007-2012 were probabilistically linked to national immunisation records. We assessed on-time coverage (defined as receipt of vaccine dose between 4 days prior to scheduled date and the recommended upper limit) for RV and compared this to diphtheria-tetanus-pertussis (DTPa) vaccine. Logistic regression modelling was used to assess independent determinants of receipt of RV.
Compared to non-Aboriginal infants, on-time RV coverage was lower for all doses among Aboriginal infants. Post the upper age limit of RV dose2, DTPa dose2 coverage increased by 9-16% to ≥90%, whereas RV coverage remained around 77% (Aboriginal) and 85% (non-Aboriginal). Compared to first-born children, the adjusted odds of receiving ≥1 RV dose if born to a mother with ≥3 previous births was 0.30 (95%CI: 0.27-0.34) among Aboriginal, and 0.53 (95%CI: 0.51-0.55) among non-Aboriginal children. Prematurity (<33 weeks), low birthweight (<1500 g), maternal age <20 years, maternal smoking during pregnancy and living in a disadvantaged area were independently associated with decreased vaccine uptake.
Aboriginal children are at greater risk of rotavirus disease than non-Aboriginal children and delayed vaccine receipt is substantially higher. Although specific programs targeting groups at risk of delayed vaccination might improve RV coverage, relaxation of upper age restrictions is most readily implementable, and its overall risk-benefit should be evaluated.

Surveys are key means of obtaining policy-relevant information not available from routine sources. Bias arising from non-participation is typically handled by applying weights derived from limited socio-demographic characteristics. This approach neither captures nor adjusts for differences in health and related behaviours between participants and non-participants within categories. We addressed non-participation bias in alcohol consumption estimates using novel methodology applied to 2003 Scottish Health Survey responses record-linked to prospective administrative data. Differences were identified in socio-demographic characteristics, alcohol-related harm (hospitalisation or mortality) and all-cause mortality between survey participants and, from unlinked administrative sources, the contemporaneous general population of Scotland. These were used to infer the number of non-participants within each subgroup defined by socio-demographics and health outcomes. Synthetic observations for non-participants were then generated, missing only alcohol consumption. Weekly alcohol consumption values among synthetic non-participants were multiply imputed under missing at random and missing not at random assumptions. Relative to estimates adjusted using previously derived weights, the obtained mean weekly alcohol intake estimates were up to 59% higher among men and 16% higher among women, depending on the assumptions imposed. This work demonstrates the universal value of multiple imputation-based methodological advancement incorporating administrative health data over routine weighting procedures.