The efficacy of an intervention can be assessed by randomizing to different diagnostic testing strategies and solving simultaneous equations: Required when assessing the effectiveness of test, trace and isolation
Year: 2018
Type: preprint
Abstract: The efficacy of an intervention can be assessed by randomising patients to different diagnostic tests instead of directly to an intervention and control. This allows the probability of an outcome to be estimated conditional on the intervention and control and thus also the odds ratio (OR) and risk reduction (RR). The tests must have different characteristics (e.g. sensitivities) with respect to the outcome. This principle is applied by allocating an intervention if the test result is positive but allocating to a control if the result is negative. This can also be done with different dichotomising thresholds for one test. The frequencies of the outcome in those with each of the four resulting observations are then used to calculate the RR and OR by solving a pair of simultaneous equations. This assumes that the RR and OR due to intervention compared to control are the same in both test groups. The calculations are illustrated by using data from a randomized controlled trial (RCT) that assessed the efficacy of an angiotensin receptor blocker (ARB) in lowering the risk of diabetic nephropathy in patients conditional on urinary albumin excretion rates (AERs). The calculations are also illustrated with simulated data for assessing the effectiveness of test, trace and isolation to reduce transmission of the SARS-Cov-2 virus by randomising to RT-PCR or LFD tests. This approach allows the sensitivity, specificity, of these tests with respect to viral spreading to be determined. The individual results of can be used to assemble curves displaying the probability of an outcome on intervention and control conditional on the measurement results by fitting spline functions. This is demonstrated using the data from the above RCT on assessing the probability of nephropathy conditional on an AER with ARB or placebo.
Source: arXiv (Cornell University)
Author: Huw Llewelyn
Institution: Aberystwyth University
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3198889657