Using observational data for personalized medicine when clinical trial evidence is limited

Abstract

Randomized clinical trials are considered the preferred approach for comparing the effects of treatments, yet data from high-quality clinical trials are often unavailable and many clinical decisions are made on the basis of evidence from observational studies. Using clinical examples about the management of infertility, we discuss how we can use observational data from large and information-rich health-care databases combined with modern epidemiological and statistical methods to learn about the effects of interventions when clinical trial evidence is unavailable or not applicable to the clinically relevant target population. When trial evidence is unavailable, we can conduct observational analyses emulating the hypothetical pragmatic target trials that would address the clinical questions of interest. When trial evidence is available but not applicable to the clinically relevant target population, we can transport inferences from trial participants to the target population using the trial data and a sample of observational data from the target population. Clinical trial emulations and transportability analyses can be coupled with methods for examining heterogeneity of treatment effects, providing a path toward personalized medicine.

Not Causal or Descriptive But Some Secret, Other Thing: Entropy as a Criterion for Causal Learning

“…Rather, all we could know was whether events tended to occur together. This correlational theory of inference was codified by Pearson (Pearl 2018, 53-91) and has remained a staple of statistical analysis: checking for associations between variables, or looking for risk factors, as in medicine (Boyko and Alderman 1990;Gershman, Guo, and Dahabreh 2018). While today's statistical education emphasizes that correlation does not equal causation, that tendency has not always been as strong among statistical practitioners.…”

Partial Versus Radical Nephrectomy for the Clinical T1a Renal Mass

“…Finally, it essential to consider heterogeneity of treatment effects; that is, that the comparative effectiveness of PN and RN varies according to patient and tumor characteristics [1,6]. In this context, the majority of studies report primarily population-averaged treatment effects [1,2,11,20].…”

“…The comparative effectiveness of partial (PN) and radical nephrectomy (RN) for the management of the small renal mass has generated great interest in contemporary urologic oncology [1][2][3]. However, the topic also represents an index case for a broader problem within the modern hierarchy of evidence-based medicine: what to do when there are limited randomized data to guide clinical decisions regarding the comparative effectiveness of competing interventions [4,5] and, perhaps even more importantly, when a limited randomized evidence base contradicts a seemingly overwhelming body of observational data [6].…”

Risk of bias in observational studies using routinely collected data of comparative effectiveness research: a meta-research study

“…They can often be resource-intensive and timeconsuming. As such, RCTs may not be able to detect effects on long-term outcomes or rare events [3][4][5]. Observational studies using routinely collected data have been used to complement RCTs [5][6][7][8].…”