There is opposite literature on the impact of waiting on health outcomes. Some studies found there is no obvious link between waiting times and health outcomes.
By comparing outcomes for people who underwent a coronary artery bypass graft (CABG), Sari et al. (2007) and Légaré et al. (2005) found that there is no difference in mortality and major adverse cardiac events between groups with different waiting times. Sari et al. (2007) compare specially patients who waited less and more than 7 days. Tuominen et al. (2010) randomised 438 patients awaiting total knee replacement patients into a short waiting time (less than 3 months) and non fixed waiting time (more than 3 months) and found that there is no difference in health-related quality-of-life. Nilsdotter & Lohmander (2002) found no differences in pre-operative status or post-operative outcome between the patients who had been on the waiting list more than and less than 3 months for unilateral total hip replacement due to osteoarthritis. Martin & Smith (1996) found no significant effect of wait times on length of stay for all surgical patients in England in 1991 and 1992. Hamilton et al. (1996) had already obtained a similar result for Canada. Hamilton et al. (2000) compare waiting times and outcomes in the US and Canada. Although waits are longer in Canada, they do not affect mortality rates. Another group of studies found opposite results. Koch et al. (1997) found poorer health outcomes with longer waits (12 weeks or more) for percutaneous transluminal coronary angioplasty. Sampalis et al. (2001) by using a prospective cohort of 266 patients from three hospitals in Montreal (Canada), found a reduction of physical functioning, vitality and other indicators for long waiters (more than 97 days). Nikolova et al. (2016) found that each additional week decrease the health gain of 0.1% after hip and knee replacement surgery. Sobolev & Fradet (2008) provided a review of the literature for CABG and suggested that long waits may worsen symptoms and clinical outcomes.Waits may also increase the probability of preoperative death (while waiting) and unplanned emergency admission (Rexius et al. , 2004 ; Sobolev et al. , 2006 ; Sobolev et al. , 2012). Hamilton et al. (2000) found that longer delays lead to significantly longer stays in the hospital after surgery for hip fractures in Quebec and Massachusetts. With all these results, the impact of waiting time on health outcomes for non-emergency treatment is less conclusive.
The reasons of divergences in these results may be due to the differences in the methodology used by these studies. Studies with no significant impact of waiting times on health outcomes are generally randomized trials, whereas studies with waiting times deteriorating health outcomes are most of all observational studies. These two groups of studies are different in the definition of waiting times, health outcomes, and the population studied. Firstly, by its conceptualisation, waiting times in randomised controlled trials are shorter (generally 1 to 3 months), which can maybe not long enough to influence the health of the patients before the admission in hospital. Secondly, observational studies use limited control variables, which may increase the probability to have significant health outcomes variables.
Finally, observational studies use more disease-specific measures of function whereas randomised controlled trials used generic health-related quality-of-life. Disease-specific measures are known to be more sensitive to assess changes within patients (Guyatt et al. , 1999 ; Puhan et al. , 2007). The contribution of our study to literature is to deal with some weaknesses of previous observational studies. In fact, our analysis focuses on real-observed-health-status measures, which are less sensitive to change in patients.
Moreover, we use a large data set with many covariates to control confounders. Finally, we use instrumental variables to deal with the potential endogeneity of waiting times.