Your example of not ethically being able to deny a patient access to CT (or any other critical diagnostic) is true for a prospective study wherein patients are divided into access and non-access groups. However, one can do a retrospective study of individuals who did or did not receive a diagnostic CT and determine morbidity and mortality. Individuals not receiving a CT can be those who went undiagnosed until autopsy or had a diagnosis but no access due to geographic location or socio-economic reasons. Retrospective studies like this are done all the time, especially to illustrate inequity of access to resources.

You could also look at the data for "Brain Attack" - Stroke, where CT evaluation is used to determine if tPA can be administered. This requires not only access to CT, but access to CT within a limited time window after the event. The time window for tPA has changed over the years, so you will have to be careful comparing studies.

Thank you for your answer, William Thoma

As I understand it can be assumed that in case of stroke the lack of CT diagnostic (i.e. lack=never until die) will lead to the death of the patient faster than its presence. Can we say that in a certain percentage of cases this will certainly happen? How much time of patient's life we can save using CT?

CT scan has been used since 1970s for cross-sectional imaging. CT scan has special place for diagnosis, treatment, and follow-up of malignancies. There has been steady progress in the efficiency of CT scan techniques. It is difficult, as well as, unethical to hold back CT scan imaging from patients who need them. Therefore, Randomized Control Trial is not a good option. However, taking CT scan as an intervention, a quasi-experimental study design may be used to show how CT scan has helped in survival of patients with malignancy. An Interrupted Time Series (ITS) study design may be used. Routinely monitored data from cancer registry may be used. Since CT scan was introduced in different parts of the world, as well as, within different parts of the country, at different times, a multiple baseline method may be used. It is important that the region/country under study should have had a cancer registry system working well in 1960-70s. The outcome to be compared should be the stage at which cancers were detected before and after CT scan was introduced. Survival rate may not be used, as it is influenced by many other factors, e.g. treatment options available.

HI Mikhail,

I have not done a comprehensive literature review for many years (most of my involvement in Brain Attack was in the late 1990s). I am not sure if anyone has done a study design to look solely at the impact on CT, but there are studies, such as: https://www.ahajournals.org/doi/10.1161/STROKEAHA.114.006570#:~:text=Table%201.%20Characteristics%20of%20IV-tPA%E2%80%93Treated%20Patients%20and%20Propensity,%20253%20%2811.8%29%20%204%20more%20rows%20 ,

designed to look at the outcomes from tPA. Someone might have done Covariance of imaging, but as mentioned by others, there are going to be multiple factors in any outcome study.