Related Experiment Videos
Using routine data to complement and enhance the results of randomised controlled trials
J D Lewsey1, A H Leyland, G D Murray
1Public Health Research Unit, University of Glasgow, UK.
Health Technology Assessment (Winchester, England)
|November 14, 2000
Summary
Routinely available hospital data can supplement randomized controlled trials (RCTs) for medical intervention evaluation. However, data limitations, such as insufficient detail, can hinder their full potential in informing clinical practice and trial design.
Area of Science:
- Health Services Research
- Clinical Epidemiology
- Medical Informatics
Background:
- Randomised controlled trials (RCTs) are the gold standard for medical intervention evaluation but face limitations in ethics, feasibility, and generalizability.
- Routinely collected health data offer potential to overcome RCT limitations, aiding trial design, outcome prediction, and real-world effectiveness assessment.
Purpose of the Study:
- To explore how routinely assembled hospital data can complement or supplement RCTs in evaluating medical interventions.
- To define contexts where RCTs are not feasible and to assess if RCT-indicated benefits are achieved in wider clinical practice using routine data.
Main Methods:
- Utilized linked Scottish morbidity and death records (1981-1995) covering 100% of acute hospital care.
- Conducted three case studies: subarachnoid hemorrhage (SAH) surgery timing and outcomes, transurethral prostatectomy (TURP) vs. open prostatectomy (OPEN) outcomes, and coronary artery bypass grafting (CABG) vs. percutaneous transluminal angioplasty (PTCA) outcomes.
Main Results:
- The SAH study indicated potential for informing trial design and evaluating new treatments, but was limited by insufficient data on symptom onset and patient condition.
- The prostatectomy study successfully addressed questions beyond RCT literature due to a large case volume, enabling subgroup analysis.
- The CABG/PTCA comparison aimed to augment RCT meta-analyses with large-scale, long-term routine data for subgroup and population-level outcome assessment.
Conclusions:
- Routinely collected data have strengths and weaknesses for evaluating medical interventions.
- Data limitations (e.g., lack of detail on patient status and timing) can prevent full realization of their potential.
- Despite limitations, routine data can address questions not feasible for RCTs and inform clinical practice and trial design when data are sufficiently comprehensive.