Friday, March 1, 2024
HomeEconomyHow does the HEOR study handle missing data? – Healthcare Economist

How does the HEOR study handle missing data? – Healthcare Economist


This is the question answered in the paper Mukherjee et al. (2023).The author defines the “HEOR study” of this article as

…a real-world evidence study using randomization for secondary/post hoc analysis
Controlled trial (RCT) data, and within-trial cost-utility analyses, where the outcome of interest is cost or PRO, including preference-based utility (e.g. EQ-5D).

The most appropriate method for imputing missing data depends on assumptions about how the data are missing:

  • Missing Completely at Random (MCAR): The observed or unobserved values ​​of all variables in the study have no effect on the probability of missing observations
  • Missing at Random (MAR). The chance of missing data for a particular variable is related to the observed value of the variable in the data set (either the observed value of other variables in the data set or the observed value of the same variable at a previous time point), but not related to the missing data. There is no way to test whether MAR holds in the data set.
  • Missing Not at Random (MNAR). In this case, the chance of missing data for a particular variable is related to the underlying value of that particular variable. MNAR can be negligible (when missing values ​​occur independently of the data collection process) or non-ignorable (when the missing mechanism has structural causes that depend on unobserved variables or the missing values ​​themselves).

To address the problem of missing data, various techniques can be used, including: complete case analysis (CCA), available case (AC) analysis, multiple imputation (MI), multiple imputation by chained equations (MICE), and predicted mean matching .

To better understand the methods commonly used in health economics and outcomes research (HEOR), the authors conducted a systematic literature review in PubMed and examined which types of statistical methods are used to account for missing costs, utility, or patient reports. outcome indicators.

The authors found that multiple imputation, chained equation multiple imputation, and complete case analysis were the most commonly used:

From 1433 identified records, 40 papers were included. Thirteen studies were economic evaluations. Thirty studies used multiple imputation, 17 of which used chained equations multiple imputation, and 15 studies used complete case analysis. Seventeen studies involved missing cost data and 23 studies involved missing outcome data. Eleven studies reported a single method, while 20 studies used multiple methods to account for missing data.

https://link.springer.com/article/10.1007/s40273-023-01297-0

The authors note that although they found a large body of HEOR methodological literature on how to deal with missing data in the context of RCTs; however, few studies have attempted to actually implement these recommendations and impute missing data.You can read the full article here.



Source link

RELATED ARTICLES

Most Popular

Recent Comments