Hi, fellow statisticians.

I know that my question seems so general, and may be similar to previous questions. However, I did not find any previous questions explicitly asking/answering my question. I am currently in a project assessing the association between BMI and mortality using dose-response meta-analysis

I discovered three types of reported hazard ratio (HR):

  • Hazard ratio in categorical values (using cut-offs) -> e.g. =30 kg/m2
  • Hazard ratio of each increment increase (for example, the authors state: "each 5 kg/m2 increase in BMI was associated with HR ...")
  • Hazard ratio of overall model (for example, the authors state: "BMI was significantly associated with earlier death (HR ...)
  • Is the interpretation of example number (2) and number (3) same? I know that the interpretation of HR value on a continuous scale is for each 1 unit increase/decrease. However, I am still clueless on the comparability of example (2) and (3).

    The reason I am asking is, I discovered some studies reporting hazard ratios by using model number (3). The authors did not state "each 1 unit increase ...", but rather "BMI was associated with ...". As I am planning to perform a dose-response meta-analysis, I have to know whether the interpretation of example number (3) is same with example number (2), in order to compare them with the generalized least squares-estimated data

    Thank you very much for the help. Any answer will be much appreciated

    Best,

    More Gilbert Lazarus's questions See All
    Similar questions and discussions