Development of novel therapies for Alzheimer’s Disease (AD) is constrained by the lack of available methods for preclinical diagnosis, despite extensive research on biomarker identification. The EMPIR NeuroMET project unites National Measurement Laboratories, clinicians and academics, to overcome limitations in measurement methods and provide a better understanding of how to improve, combine and analyse measurements in AD diagnosis and treatment. Comparability through SI (System of International Units) traceability and uncertainty analysis is an, as yet, unmet requirement for regulatory approval of biomarkers, patient centred outcome measures, clinical thresholds and new therapeutic drugs. We will report on:

• Multimodal statistical analysis on blood, CSF and saliva biomarkers data from the NeuroMET cohort generated by using mass spectroscopy and immunoassay platforms, including a novel immunoassay approach to overcome matrix effects when relative quantification is not sensitive enough. A new digital PCR approach was developed to assess microRNAs quantities in blood to compare with established biomarkers.
• Progress towards the development of mass spectrometry reference measurement procedures traceable to the SI for t-tau and α-synuclein in cerebrospinal fluids.
• Development of ultrahigh field Magnetic Resonance Imaging and Spectroscopy protocols for increased spatial and spectral resolution and decreased uncertainty, and their application to the NeuroMET cohort.
• Improved cognitive assessment protocols, with improved metrological evaluation of cognitive performance scores and the development of construct specification equations for various cognitive protocols and biomarkers.
• Potential relationships between volumes of AD-related brain structures and neurometabolite concentrations with measured cognitive function.

Improved reference methods to underpin the production of calibrators and improve measurement comparability of established biomarkers has the potential to further the understanding of AD and boost research for disease modifying therapies.

In establishing metrological quality assurance in healthcare, one major hurdle is the correct treatment ofordinal data typical of questionnaires, performance tests and other categorical data collected widely in care.Despite being known well over a century, there are still many examples of measurements in healthcare – forexample, (i) on-line tables of percentage performance indicators (e.g. fraction of patients seeing a doctorwithin seven days) and (ii) correlation plots for Alzheimer sufferers of cognitive scores against biomarkerconcentration – where the ‘counted fraction’ distortion of scales is not compensated for. The Rasch form ofgeneralised linear model not only handles counted fraction ordinality but also enables separation of objectand instrument attributes (such as task difficulty and patient ability) essential for metrological restitutionin measurement systems in healthcare. A perspective is given of a new kind of certified reference materialemploying causal Rasch models in terms of construct specification equations for metrological item bankingin the social sciences. This is part of the response to a recent call for: ‘a new international body to bringtogether metrology, psychometrics, philosophy, and clinical management to support the global comparabilityand equivalence of measurement results in patient centred outcome measurement to improve healthcare’.

Alongside a lack clinically validated, minimally invasive diagnostic tools for early diagnosis and/or monitoring of disease progression in Alzheimer’s Disease (AD), there are no sound metrological assessment protocols and measurements for cognition, nor measurement comparability through SI (System of International Units), traceability and uncertainty for regulatory approval of biomarkers. As part of the on-going EMPIR HLT04 NeuroMET project in which several national metrology institutes work together with clinicians and academics to overcome specific measurement issues to improve diagnosis and disease progression, we describe here the justification for and development of the ‘NeuroMET Memory Metric’ (version 0.1).

Re-examination of traditional widely used ‘legacy’ cognitive assessment protocols using invariant measurement theory aims at more accurately capturing the patient’s cognitive ability and improving the analysis of correlations with various AD biomarkers. Two principal elements provide sound metrological underpinnings: (i) a correct formulation of a measurement model; and (ii) proper handling of the ordinal cognitive data. In turn, this enables formulation of novel construct specification equations for patient cognitive ability as a function of diverse biomarkers (e.g., in plasma, CSF and saliva together with MRI/MRS data) as well as for cognitive task difficult as a function of test design.

To further improve the accuracy in patient’s cognitive ability work is now in progress to develop a NeuroMET Memory Metric based on legacy cognitive assessments (e.g., MMSE, Corsi Block Test, Digital Span Test). This work can be ascribed a level-5 construct theory. This means the realization of more fit-for-purpose, better targeted and better administered cognitive measurement systems. It will also enable traceable calibration of both additional cognitive tasks as well as the effects of intervention (or disease progression) on the cognitive ability of each individual patient.

Did you ever wonder how to account for biases such as ceiling eects in your (cognitive) data?

Metrological quality assurance of human-based responses is in its infancy and analyzing categorial data and other human responses is challenging. However, there is a need to tackle those challenges to ensure that decisions about health care are made correctly. Quality assured comparability, interoperability and decision-making can successfully be done by applying sound metrological approaches to enable traceability as well as stressing declaration of measurement uncertainties. In the seminar, we present metrological approaches to ensure quality assurance of categorical data,such as cognitive assessments and other human-reported responses. This is followed by a hands-onworkshop where you are welcome to bring your own or freely available data for analyses.

Objective: To evaluate the Patient Participation in Rehabilitation Questionnaire (PPRQ) according to Rasch measurement theory. Method: Five hundred twenty-two post-discharge patients from a neurological rehabilitation unit were included. The PPRQ questionnaire comprises 20 items rated by a cohort of 522 patients about their experiences of participating in rehabilitation. The measurement properties of the PPRQ were evaluated by Rasch analysis of the responses. Results: The Rasch analysis of 20 items showed some major misfits, particularly three items addressing the involvement of family members. After removing those items, the model fit improved and no significant DIF remained. Despite improvements, person values (−2.96 to 4.86 logits) were not fully matched by the item values (−0.61 to 0.77 logits). Neither did the t test for unidimensionality meet the criterion of 5%, and local dependency was present. The unidimensionality and local dependency could, however, be accommodated for by four testlets. Conclusion: The PPRQ-17 showed that a ruler with a reasonable and clinical hierarchy can be constructed, although the expectations of dimensionality and local dependency need to be evaluated further. Despite room for further development, PPRQ-17 nevertheless shows improved measurement precision in terms of patient leniency compared with previous evaluations with classical test theory. In turn, this can play a crucial role when comparing different rehabilitation programs and planning tailored care development activities.

Patient centered outcomes pertain to a patient's beliefs, opinions and needs in conjunction with a clinician's medical expertise and assessment. The rise of patient-centered outcome (PCO) measurement parallels increased interest in patient-centered care. PCO measures offer the opportunity for more meaningful measurement of health outcomes informative enough to guide treatment decisions. However, it has been suggested that, for practical and scientific reasons, existing PCO measures are currently not capable of delivering the kind of quality assured measurement required for high-stakes decision making. Potential solutions include: addressing the lack of units in PCO measurement through recourse to mathematical models devised to define meaningful, invariant, and additive units of measurement with known uncertainties; establishing coordinated international networks of key stakeholders guided by five principles (i.e., collaboration, alignment, integration, innovation and communication); better use of technology leveraging measurement through item banks linking PCO reports via common items, common patients, or specification equations based in strong explanatory theory. And finally ensuring PCO measurement always is associated with: (1) a clear definition of the measurand in regards to the intended clinical use; (2) a clear definition of the clinically allowable error of measurement; (3) international cooperation and consensus to navigate the complexities of the development of metrologically sound reference measurement systems; and (4) continued clinical validation of newly calibrated measures. In this article, we illustrate the principles to improve PCO measures with examples from breast cancer, vision-related patient-reported outcome measures, and dementia clinician-reported and performance outcome measures.

Is it realistic to aspire to the same kind of quality-assurance of measurement in person-centred care, currently being implemented in healthcare globally, as is established in the physical sciences and engineering? Ensuring metrological comparability (‘traceability’) and reliably declaring measurement uncertainty when assessing patient ability or increased social capital are however challenging for subjective measurements often characterised by large dispersion. Drawing simple analogies between ‘instruments’ in the social sciences – questionnaires, ability tests, etc.–  and  engineering instruments such as thermometers does not go far enough. A possible way forward apparently equally applicable to both physical and social measurement, seems to be to model inferences in terms performance metrics of a measurement system. Person-centred care needs person-centred measurement and a full picture of the measurement process when Man acts as a measurement instrument is given in the present paper. This complements previous work by presenting the process, step by step, from the observed indication (e.g. probability of success, P_success, of achieving atask), through restitution with Rasch Measurement Theory, to the measurand (e.g. task difficulty). Rasch invariant measure theory can yield quantities –‘latent’ (or ‘explanatory’) variables such as task challenge or person ability – with characteristics akin to those of physical quantities. Metrological references for comparability via traceability and reliable estimates ofuncertainty and decision risks are then in reach even for perceptive measurements (and other qualitative properties). As a case study, the person-centred measurement of cognitive ability is examined, as part of the EUproject EMPIR 15HLT04 NeuroMet, for Alzheimer’s, where better analysis of correlations with brain atrophy is enabled thanks to the Rasch metrological approach.

The development of novel therapies for Alzheimer’s Disease (AD) is constrained by the lack of available methods for preclinical diagnosis, despite extensive research on biomarker identification. Here, we present an update of progress from EMPIR NeuroMET, a project combining diverse expertise from five National Measurement Institutes (NMIs), with clinicians and academics, to overcome limitations in measurement methods in neurodegenerative disease diagnosis and treatment.

Measurements in the social sciences – with ‘instruments’ such as questionnaires, ability tests, – in education, healthcare and so on, need metrological quality assurance. A patient, for instance, expects the same quality of care wherever and whenever provided. This is a challenge since the usual tools of statistics do not always work on the categorical scales typical of such measurements. Modelling a measurement system where the instrument is a human being, and where the output is a performance metric, i.e., how well the set-up performs an assessment, appears to be a way forward. This BEMC Colloquium will present the necessary tools, such as psychometric Rasch measurement theory, and will be followed by a hands-on workshop where you yourselves can analyse cases such as (i) the Quality of the BEMC Colloquium Series and (ii) a Healthy Lifestyle.

Measurement units have historically been defined as quantities (i.e., specific properties) of objects, such as the mass of a particular piece of metal or the length of a particular rod. While the current International System of Units (SI) Brochure endorses this position, the draft 9th SI Brochure proposes to change it, and instead define measurement units as values of quantities. The reason for this proposed change is not provided, but it does not seem plausible that it is related to the redefinition of the SI units in terms of fundamental constants of nature: the very concept of what a unit is does not depend on the concrete way any given unit is defined. This paper is intended to open a discussion of whether measurement units should be defined as quantities or as quantity values, and provides our rationale for maintaining the definition of units as quantities.