skip to content W3C SVG

Open Tabs (01-31-21): Propensity Scores and xaringan

Publish date: Feb 5, 2021
Last updated: Feb 5, 2021
Tags:

Memes

👇 This is me whenever I need to share my screen…

Propensity Scores

  • I attended Todd Wagner’s talk on Propensity Scores as part of the VA HERC lecture series. See slides.
  • How to use Propensity scores:
    • Compare individuals based on similar PS scores (a matched analysis)
    • Conduct subgroup analyses on similar groups (stratification)
    • Include the PS as a covariate (quintiles of the PS) in the regression model
    • Use the PS to weight the regression (i.e., place more weight on similar cases)
    • Use both 3 and 4 together (doubly robust)
  • He stressed that the choice of matching estimator (for Propensity Score Matching) is extremely important.
  • He also emphasized an important limitation of PSM: observed variables are used for PSM, and those unobserved variables can be problematic as a possible source of confounding.
  • Compare the balance using Love plots (“is there balance after matching?")
  • Papers he referenced that I’ve added to my reading list:
    • Garrido MM, Kelley AS, Paris J, et al. Methods for constructing and assessing propensity scores. Health Serv Res. 2014;49(5):1701-1720. doi:10.1111/1475-6773.12182
    • Brooks JM, Ohsfeldt RL. Squeezing the balloon: propensity scores and unmeasured covariate balance. Health Serv Res. 2013;48(4):1487-1507. doi:10.1111/1475-6773.12020
    • King G, Nielsen R. Why propensity scores should not be used for matching. Polit Anal. 2019;27(4):435-454. doi:10.1017/pan.2019.11
  • Relevant R packages:
    • MatchIt (site, GitHub)
      • “a simple and straightforward interface to various methods of matching for covariate balance in observational studies”
      • Matching methods include “nearest neighbor matching, optimal pair matching, optimal full matching, genetic matching, exact matching, coarsened exact matching, and subclassification”
    • WeightIt (site, GitHub)
      • “a one-stop package to generate balancing weights for point and longitudinal treatments in observational studies”

Healthcare

  • 🎉 Reimbursement requirements are changing and doctors are celebrating the possibility of reduced note bloat and superfluous or unnecessary documentation. Jeff Linder of NMH has even shared a nice template of what the future may have in store.
    • “The new guidelines, scheduled to take effect January 1, 2021, focus on medical decision-making, and are intended to more closely reflect the actual work providers perform in treating patients’ conditions. The goal is that providers have more time to spend on patient care with less required for documentation or other administrative processes. The desired result is higher quality documentation leading to higher quality patient care.” (Source)
Jeff Linder's Note Template following changes to CMS documentation requirements.

Figure 1: Jeff Linder's Note Template following changes to CMS documentation requirements.

Papers

R goodness

  • xaringancolor by Emil Hvitfeldt (post)
    • “allows you to specify shared text colors in the text, equations and code sections of xaringan slides”
    • This joins a growing list of xaringan goodness: xaringanExtra and xaringanthemer (both by Garrick Aden-Buie)
  • And even more xaringan-iness: Convert your PowerPoint presentations to HTML (Karandeep Singh)
  • Make your Shiny apps look like a Nintendo throwback using bslib and thematic (Nick Strayer)
library(bslib)

bs_theme(
  base_font=font_google("Press Start 2P"), 
  bg="#e7e6e1", fg="#9e252a") %>%
  bs_add_rules(
    '@import 
    "unpkg.com/nes.css@latest/css/nes.min.css"')
thematic::thematic_shiny(font="auto")
  • details (site, GitHub)
    • I even created a custom VSCode snippet to create toggles in my markdown files this week 👇
// markdown.json.code-snippets
	"Toggled content": {
		"prefix": "toggle",
		"body": [
		  "<details>",
		  "<summary></summary><p>",
		  "",
		  "</p>",
		  "</details>"
		],
		"description": "Toggled content"
	  }