Selecting measurement instruments (questionnaires)

Aim

  • To search for existing measurement instruments
  • To choose the most appropriate measurement instrument(s)

 

Requirements

  • Short conclusion explaining the final decision making of the questionnaire(s)
  • Validity and reliability of the chosen questionnaire(s) should be known

 

Documentation

  • Short conclusion about the final decision making
  • List of considered questionnaire(s) with their validity and reliability

 

Responsibilities

Executing researcher: Providing documentation why (a) certain questionnaire(s) was/were chosen
Project leaders: Knowledge of the final decision making of the questionnaire(s)
Research assistant: N.a.

 

How To

Define the construct and conceptual model
The construct of interest indicates what kind of measurement instrument you should select, e.g. a self-reported or interview-administered questionnaire, a performance-based test, or a clinical measure. For example, when you want to measure physical functioning, you could be interested in what people think they can do, what they actually do (e.g. they may not walk stairs because there is always an elevator to use), or what they can do (e.g. under supervision of a physiotherapist). Although these constructs are closely related, they are not the same. Models like the International Classification of Functioning (ICF) (WHO, 2011) or the model by Wilson and Cleary (1995) can be used as a conceptual model to relate the construct in which you are interested in to other related and unrelated constructs.

 

Define measurement aim
The aim of the measurement has consequences for which measurement properties are most important. A measurement instrument may be applied as a discriminative instrument, i.e. to measure differences between people at a single point in time, as an evaluative instrument, i.e. to measure changes within people over time, or as a predictive instrument, i.e. to classify individuals according to their prognosis.

 

Define target population
Measurement instruments, or items within an instrument may be appropriate for a specific target population. Therefore, consider thoroughly whether the instrument is appropriate for patients’ age, gender, diagnosis, severity of the disease and setting.

 

Feasibility
Take issues into account such as costs, duration, required expertise of the target population, and respondent burden.

 

Measurement properties
Validity and interpretability are always important. When an instrument is used in a discriminative application, reliability is important, and when an instrument is used in an evaluative application, measurement error and responsiveness are important. When instruments are used to detect very small changes, or on individual level, it is important that the measurement error of the instrument is extremely small. Internal consistency is important for multi-item (sub) scales measuring uni-dimensional constructs.

 

Appendices/references/links

  • See also cosmin.nl for an overview and definitions of all important measurement properties, and a checklist to assess the methodological quality of studies on measurement properties
  • See also kmin-vumc.nl to find a number of links to electronic databases and websites that contain measurement instruments or systematic reviews on measurement instrument.
    1. Use Pubmed when no review for your construct of interest is available with search filter Terwee et al. 2009.
    2. Use De Vet et al. 2011 for how to perform a systematic review on measurement properties of measurement instruments.
    3. When the best instrument is not available in Dutch, it needs to be translated using a forward and backward method. We refer to a guideline available (in Dutch) at kmin-vumc.nl, or the guidelines proposed by Beaton et al. 2000.
    4. When selecting PROs and/or PROMs, then take a look in the NFU guide for selecting PROs and PROMs (Click here for a Dutch version).

 

Audit questions

  1. Are the concepts that are intended to measure clearly described a priori in the study protocol?
  2. Has a systematic literature review been carried out in order to find the most appropriate questionnaire per concept?
  3. Is there sufficient evidence to support the selection of the questionnaires used?
    1. Have the measurement properties been evaluated?

 

V3.0: 13 Oct 2016: Revision guideline
V2.1: 12 May 2015: Revision format
V2.0: 27 May 2011: Guideline entirely rewritten and divided in 3 guidelines
V1.0 1 Jan 2010: Translation into English and updated
V1.2: 2 March 2007: Point added about questionnaires copyright
V1.1: 13 February 2006: References updated, text improved in various places, sections adapted
V1.0: This guideline has been rewritten entirely