Biospecimens are the essential substrates for human biomarker research. Across the globe, biobanks have developed the facilities and mechanisms to collect, process, store and distribute those substrates to researchers. However, despite some notable successes, less than one hundred of the tens of thousands of purported biomarkers have been independently validated. We propose the need for a new paradigm in biobanking; simply pursuing larger numbers of participants, larger networks of biobanks and higher sample integrity will not, in itself, transform the success rate or efficiency of biomarker research. We propose that biobanks must embrace the intrinsic observational nature of biospecimens and furnish the recipients of biospecimens with the population metrics (descriptive statistics) that can facilitate the scientific rigor that is mandated in other areas of observational research. In addition, we discuss the value of population-based ascertainment and recruitment and the importance of the timing of biospecimen collections. Any assessment of biospecimen quality must go beyond the sample itself and consider both the patient/participant selection and the most appropriate and informative timing for specimen collection, particularly prior to any treatment intervention in diseased populations. The examples and rationales that we present are based largely on cancer-related collections because the feasibility of population metrics is greatly assisted by the comprehensive registries that are more common for cancer than other chronic diseases. Changing the biobanking paradigm from tacitly 'experimental' to explicitly 'observational' represents a profound but urgent methodological shift that will influence the establishment, management, reporting and impact of biobanks in the twenty-first century.
Keywords: Biobanking quality; Biomarker research; Biospecimen; Cancer; Epidemiology; Networking and integration; Observational; Population-based; Registries; Selection bias; Validation.