Analytical Methods for Observational Data to Generate Hypotheses and Inform Clinical Decisions

Todd DeWees, Carlos E. Vargas, Michael A. Golafshar, William Scott Harmsen, Amylou Dueck

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Randomized controlled trials have been considered the gold standard in informing clinical decision-making while observational studies have generally been utilized to generate hypotheses for future studies. The rising cost of randomized studies along with increased difficulty in accrual has led the clinical community to consider utilizing observational studies to inform clinical decisions. Various statistical methods exist to analyze observational data. Researchers must consider each method carefully, paying specific attention to its ability to answer the hypotheses, while ensuring the underlying assumptions are met. While each has its own strengths and weaknesses, research has shown that each method may yield similar estimates of treatment effect when conducted appropriately. We describe several commonly used analytical methods including their: strengths, weaknesses, and common missteps in order to inform and serve as a reference to the broader oncology community.

Original languageEnglish (US)
JournalSeminars in Radiation Oncology
DOIs
StatePublished - Jan 1 2019

Fingerprint

Observational Studies
Randomized Controlled Trials
Research Personnel
Costs and Cost Analysis
Research
Clinical Decision-Making

ASJC Scopus subject areas

  • Oncology
  • Radiology Nuclear Medicine and imaging
  • Cancer Research

Cite this

Analytical Methods for Observational Data to Generate Hypotheses and Inform Clinical Decisions. / DeWees, Todd; Vargas, Carlos E.; Golafshar, Michael A.; Harmsen, William Scott; Dueck, Amylou.

In: Seminars in Radiation Oncology, 01.01.2019.

Research output: Contribution to journalArticle

@article{8ba7af617aee43138df95429609c1cd7,
title = "Analytical Methods for Observational Data to Generate Hypotheses and Inform Clinical Decisions",
abstract = "Randomized controlled trials have been considered the gold standard in informing clinical decision-making while observational studies have generally been utilized to generate hypotheses for future studies. The rising cost of randomized studies along with increased difficulty in accrual has led the clinical community to consider utilizing observational studies to inform clinical decisions. Various statistical methods exist to analyze observational data. Researchers must consider each method carefully, paying specific attention to its ability to answer the hypotheses, while ensuring the underlying assumptions are met. While each has its own strengths and weaknesses, research has shown that each method may yield similar estimates of treatment effect when conducted appropriately. We describe several commonly used analytical methods including their: strengths, weaknesses, and common missteps in order to inform and serve as a reference to the broader oncology community.",
author = "Todd DeWees and Vargas, {Carlos E.} and Golafshar, {Michael A.} and Harmsen, {William Scott} and Amylou Dueck",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.semradonc.2019.05.003",
language = "English (US)",
journal = "Seminars in Radiation Oncology",
issn = "1053-4296",
publisher = "W.B. Saunders Ltd",

}

TY - JOUR

T1 - Analytical Methods for Observational Data to Generate Hypotheses and Inform Clinical Decisions

AU - DeWees, Todd

AU - Vargas, Carlos E.

AU - Golafshar, Michael A.

AU - Harmsen, William Scott

AU - Dueck, Amylou

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Randomized controlled trials have been considered the gold standard in informing clinical decision-making while observational studies have generally been utilized to generate hypotheses for future studies. The rising cost of randomized studies along with increased difficulty in accrual has led the clinical community to consider utilizing observational studies to inform clinical decisions. Various statistical methods exist to analyze observational data. Researchers must consider each method carefully, paying specific attention to its ability to answer the hypotheses, while ensuring the underlying assumptions are met. While each has its own strengths and weaknesses, research has shown that each method may yield similar estimates of treatment effect when conducted appropriately. We describe several commonly used analytical methods including their: strengths, weaknesses, and common missteps in order to inform and serve as a reference to the broader oncology community.

AB - Randomized controlled trials have been considered the gold standard in informing clinical decision-making while observational studies have generally been utilized to generate hypotheses for future studies. The rising cost of randomized studies along with increased difficulty in accrual has led the clinical community to consider utilizing observational studies to inform clinical decisions. Various statistical methods exist to analyze observational data. Researchers must consider each method carefully, paying specific attention to its ability to answer the hypotheses, while ensuring the underlying assumptions are met. While each has its own strengths and weaknesses, research has shown that each method may yield similar estimates of treatment effect when conducted appropriately. We describe several commonly used analytical methods including their: strengths, weaknesses, and common missteps in order to inform and serve as a reference to the broader oncology community.

UR - http://www.scopus.com/inward/record.url?scp=85067194446&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85067194446&partnerID=8YFLogxK

U2 - 10.1016/j.semradonc.2019.05.003

DO - 10.1016/j.semradonc.2019.05.003

M3 - Article

JO - Seminars in Radiation Oncology

JF - Seminars in Radiation Oncology

SN - 1053-4296

ER -