The population-attributable fraction for time-dependent exposures and competing risks-A discussion on estimands

Dealing with time-dependent exposures and confounding when defining and estimating attributable fractions-Revisiting estimands and estimators

The population-attributable fraction (PAF) is commonly interpreted as the proportion of events that can be ascribed to a certain exposure in a certain population. Its estimation is sensitive to common forms of time-dependent bias in the face of a time-dependent exposure. Predominant estimation approaches based on multistate modeling fail to fully eliminate such bias and, as a result, do not permit a causal interpretation, even in the absence of confounding. While recently proposed multistate modeling approaches can successfully eliminate residual time-dependent bias, and moreover succeed to adjust for time-dependent confounding by means of inverse probability of censoring weighting, inadequate application, and misinterpretation prevails in the medical literature. In this paper, we therefore revisit recent work on previously proposed PAF estimands and estimators in settings with time-dependent exposures and competing events and extend this work in several ways. First, we critically revisit the interpretation and applied terminology of these estimands. Second, we further formalize the assumptions under which a causally interpretable PAF estimand can be identified and provide analogous weighting-based representations of the identifying functionals of other proposed estimands. This representation aims to enhance the applied statistician's understanding of different sources of bias that may arise when the aim is to obtain a valid estimate of a causally interpretable PAF. To illustrate and compare these representations, we present a real-life application to observational data from the Ghent University Hospital ICUs to estimate the fraction of ICU deaths attributable to hospital-acquired infections.

The population-attributable fraction for time-to-event data

Background

Even though the population-attributable fraction (PAF) is a well-established metric, it is often incorrectly estimated or interpreted not only in clinical application, but also in statistical research articles. The risk of bias is especially high in more complex time-to-event data settings.

Methods

We explain how the PAF can be defined, identified and estimated in time-to-event settings with competing risks and time-dependent exposures. By using multi-state methodology and inverse probability weighting, we demonstrate how to reduce or completely avoid severe types of biases including competing risks bias, immortal time bias and confounding due to both baseline and time-varying patient characteristics.

Results

The method is exemplarily applied to a real data set. Moreover, we estimate the number of deaths that were attributable to ventilator-associated pneumonia in France in the year 2016. The example demonstrates how, under certain simplifying assumptions, PAF estimates can be extrapolated to a target population of interest.

Conclusions

Defining and estimating the PAF in advanced time-to-event settings within a framework that unifies causal and multi-state modelling enables to tackle common sources of bias and allows straightforward implementation with standard software packages.

Attributable Mortality of Ventilator-associated Pneumonia. Replicating Findings, Revisiting Methods

Rationale: Estimating the impact of ventilator-associated pneumonia (VAP) from routinely collected intensive care unit (ICU) data is methodologically challenging.

Objectives: We aim to replicate earlier findings of limited VAP-attributable ICU mortality in an independent cohort. By refining statistical analyses, we gradually tackle different sources of bias.

Methods: Records of 2,720 adult patients admitted to Ghent University Hospital ICUs (2013–2017) and receiving mechanical ventilation within 48 hours after admission were extracted from linked Intensive Care Information System and Computer-based Surveillance and Alerting of Nosocomial Infections, Antimicrobial Resistance, and Antibiotic Consumption in the ICU databases. The VAP-attributable fraction of ICU mortality was estimated using a competing risk analysis that is restricted to VAP-free patients (approach 1), accounts for VAP onset by treating it as either a competing (approach 2) or censoring event (approach 3), or additionally adjusts for time-dependent confounding via inverse probability weighting (approach 4).

Results: A total of 210 patients (7.7%) acquired VAP. Based on benchmark approach 4, we estimated that (compared with current preventive measures) hypothetical eradication of VAP would lead to a relative ICU mortality reduction of 1.7% (95% confidence interval, −1.3 to 4.6) by Day 10 and of 3.6% (95% confidence interval, 0.7 to 6.5) by Day 60. Approaches 1–3 produced estimates ranging from −0.7% to 2.5% by Day 10 and from 5.2% to 5.5% by Day 60.

Conclusions: In line with previous studies using appropriate methodology, we found limited VAP-attributable ICU mortality given current state-of-the-art VAP prevention measures. Our study illustrates that inappropriate accounting of the time dependency of exposure and confounding of its effects may misleadingly suggest protective effects of early-onset VAP and systematically overestimate attributable mortality.

Predicting Potential Prevention Effects on Hospital Burden of Nosocomial Infections: A Multistate Modeling Approach

OBJECTIVES

Hospital-acquired infections (HAIs) place a substantial burden on health systems. Tools are required to quantify the change in this burden as a result of a preventive intervention. We aim to estimate how much a reduction in the rate of hospital-acquired infections translates into a change in hospital mortality and length of stay.

METHODS

Using multistate modelling and competing risks methodology, we created a tool to estimate the reduction in burden after the introduction of a preventive effect on the infection rate. The tool requires as inputs the patients' length of hospital stay, patients' infection information (status, time), patients' final outcome (discharged alive, dead), and a preventive effect. We demonstrated the methods on both simulated data and 3 published data sets from Germany, France, and Spain.

RESULTS

A hypothetical prevention that cuts the infection rate in half would result in 21 lives and 2212 patient-days saved in French ventilator-associated pneumonia data, 61 lives and 3125 patient-days saved in Spanish nosocomial infection data, and 20 lives and 1585 patient-days saved in German nosocomial pneumonia data.

CONCLUSIONS

Our tool provides a quick and easy means of acquiring an impression of the impact a preventive measure would have on the burden of an infection. The tool requires quantities routinely collected and computation can be done with a calculator. R code is provided for researchers to determine the burden in various settings with various effects. Furthermore, cost data can be used to get the financial benefit of the reduction in burden.

Population attributable fractions of mortality in people living with HIV: roles of delayed antiretroviral therapy, hepatitis coinfections and social factors

OBJECTIVES

Despite free access to antiretroviral therapy (ART) from 1996 onward, and treatment for all people living with HIV (PLWHIV) from 2013, mortality in Brazil has not homogeneously decreased. We investigated to what extent delayed ART, hepatitis coinfections and sociodemographic factors predict all-cause mortality in Brazilian PLWHIV.

DESIGN

We included PLWHIV at least 18 years, with complete CD4 cell count data, followed up between 2007 and 2015 in Brazil.

METHODS

After multiple imputation, an extended Cox model helped estimate the effects of fixed and time-varying covariates on mortality.

RESULTS

The study population (n = 411 028) were mainly male (61%), white (55%), 40 years or less (61%), heterosexually HIV infected (71%), living in the Southeast region (48%) and had basic education (79%). Hepatitis C virus and hepatitis B virus coinfection prevalences were 2.5 and 1.4%, respectively. During a 4-year median follow-up, 61 630 deaths occurred and the mortality rate was 3.45 (95% confidence interval: 3.42-3.47) per 100 person-years. Older age, male sex, non-white ethnicity, illiteracy/basic education and living outside the Southeast and Central-West regions were independently associated with increased mortality. The main modifiable predictors of mortality were delayed ART (i.e. CD4 cell count <200 cells/μl at ART initiation) (adjusted population attributable fraction: 14.20% [95% confidence interval: 13.81-14.59]), being ART-untreated (14.06% [13.54-14.59]) and ART-treated with unrecorded CD4 at ART initiation (5.74% [5.26-6.21]). Hepatitis C virus and hepatitis B virus coinfections accounted for 2.44 [2.26-2.62] and 0.42% [0.31-0.53] of mortality, respectively.

CONCLUSION

The current study demonstrates that besides early ART and coinfection control, actions targeting males, non-whites and illiterate people and those with basic education are important to reduce avoidable deaths among Brazilian PLWHIV.

Quantification and interpretation of attributable mortality in core clinical infectious disease journals

Attributable mortality is an important metric that mirrors the public health effect of a potentially harmful infection by accounting not only for the effect of infection on mortality, but also for its prevalence within the target population. We did a systematic literature review to identify how attributable deaths were quantified and interpreted in core clinical infectious diseases journals between Jan 1, 2013, and April 6, 2020. Of the 1591 abstracts screened, 234 entered the primary analysis. Our summary of the epidemiological measures used in these articles reveals fundamental shortcomings in the conception of attributable mortality. Because of its importance as a basis for decision making on public health matters, it is essential to correctly quantify and report on attributable mortality. Our recommendation for quantification and reporting of attributable deaths aids clinical researchers in the correct statistical assessment of the burden of infections. Fictional as well as real data is used to illustrate these issues.