Odds ratio and hazard ratio - why does this paper use both? Am trying to get my head around the difference between hazard and odds ratio, and was looking at the practical example given in this paper: http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(15)01126-5/abstract
More specifically this section:

Between May 1, 2010, and Nov 14, 2011, 2507 patients were recruited to the study and randomly assigned to closure of the mesenteric defects (n=1259) or non-closure (n=1248). 2503 (99·8%) patients had follow-up for severe postoperative complications at day 30 and 2482 (99·0%) patients had follow-up for reoperation due to small bowel obstruction at 25 months. At 3 years after surgery, the cumulative incidence of reoperation because of small bowel obstruction was significantly reduced in the closure group (cumulative probability 0·055 for closure vs 0·102 for non-closure, hazard ratio 0·56, 95% CI 0·41–0·76, p=0·0002). Closure of mesenteric defects increased the risk for severe postoperative complications (54 [4·3%] for closure vs 35 [2·8%] for non-closure, odds ratio 1·55, 95% CI 1·01–2·39, p=0·044), mainly because of kinking of the jejunojejunostomy.

Why does the analysis of reoperation due to small bowel obstruction use hazard ratio, and the analysis of severe postoperative complications use odds ratio? They both seem to be measuring a consequence of the factor under investigation, albeit at different endpoints. 
 A: The assessment of severe postoperative complications was done at a fixed time (30 days) after surgery. This is a standard time point at which to evaluate short-term complications of surgical procedures. Almost all patients received follow-up at that time, and there was apparently no interest in how quickly the complications developed within those 30 days after surgery. So the issue is a simple yes/no for complications by 30 days, and the odds ratio is an appropriate comparison.
Re-operation, in contrast, could occur at any time during the study, years after the initial surgery. Although some cases had 3 or more years of follow up at the end of the study, many had shorter follow up. In that type of situation, with varying lengths of individual follow up and with interest in how quickly re-operation might occur, survival analysis and assessment of hazard ratios is a useful way to proceed. It allows inclusion of all available data on the study participants without setting an arbitrary follow-up cutoff time. An arbitrary cutoff time would necessarily exclude participants with shorter follow up and would exclude information on re-operations that occurred after the cutoff time.
