Our methods

Study identification and data extraction at a glance

Study identification and data extraction at a glance

Is the report a primary study?

Primary studies may use a longitudinal or cross-sectional design to calculate an MID. 

Do patients (of any age, disease, or condition given any intervention) respond to a PROM in the context of the study?

A PROM is a measure of a patient’s health condition reported by the patient, without amendment or interpretation by a clinician or anyone else. PROMs may measure health-related quality of life, functional ability, symptom severity, psychological distress, or well-being. At times, patients may be incapable of responding to a PRO; therefore, the PROM may be completed by a proxy, such as a parent, guardian, or other family member. We do not consider functional outcome measures, such as the six-minute walk test, disease activity scores, or any other “objective” measure of health status a PROM.

Do authors calculate an anchor-based MID?

An MID may be calculated using a distribution-based or anchor-based method. When using distribution-based methods for calculating MIDs (e.g., 0.5 standard deviation method), researchers rely on the statistical characteristics of the study sample. However, when using anchor-based methods to calculate an MID, researchers related differences in PROM scores to an external criterion such as a transition rating, a clinical endpoint, or other PROM.

Our rigorous five-step process

  • Study identification

    We systematically search
    We systematically search
    • Medline
    • EMBASE
    • PsycINFO
    • CINAHL
    • PROQOLID
    We use keywords such as
    We use keywords such as
    • minimal important difference
    • clinically important difference
    • health outcome assessment
    • health status indicator
    • severity of illness index
    Eligibility criteria
    Eligibility criteria
    1. Is the report a primary study?
    2. Are the authors estimating an MID for a PROM?
    3. Is the MID derived using an anchor-based method?
    We include any anchor-based MID irrespective of the patients’ condition or disease, type of intervention used in the study, or nature of the anchor.
  • Screening

    Two screeners independently assess the eligibility of each reference based on titles and abstracts.
    Two screeners independently examine the full text of all studies deemed relevant based on title and abstract screening.
    Any discrepancies in eligibility are resolved by one or more senior reviewers.
  • Data extraction

    Two reviewers independently extract data from each study.
    The following information is recorded:
    • geographic location of the study
    • participant demographics (age, sex, description of clinical condition)
    • description of any interventions administered
    • description of the PROM (construct(s), domain (if applicable), and scale)
    • description of the anchor (type, construct(s), range of options/categories/values, threshold selected to represents a “small but important difference” and calculate the MID, and who completes the anchor)
    • details about the MID estimation (follow-up length (in longitudinal studies) or timepoint at which PROM and anchor were evaluated (in cross-sectional studies), analytical method, whether the MID is reported in absolute or relative terms, whether the MID reflects improvement, worsening or both improvement and worsening)
    • MID estimate and associated precision, and the number of participants included in the MID calculation
  • Credibility Assessment

    We perform a credibility assessment for each MID. Two reviewers independently judge whether the MID satisfies each credibility criterion by selecting one of five response options.
    There are five key credibility criteria:
    1.
    Did the patient (or proxy) respond directly to both the PROM and the anchor?
    2.
    Is the anchor easily understandable and relevant for patients?
    3.
    Has the anchor shown good correlation with the PROM?
    4.
    Is the MID precise?
    5.
    Does the threshold or difference between groups on the anchor used to estimate the MID reflect a small but important difference?
    There are four additional credibility criteria a transition rating anchor is used:
    1.
    Is the amount of elapsed time between baseline and follow-up measurement for MID estimation optimal?
    2.
    Does the transition item have a satisfactory correlation with the PROM score at follow-up?
    3.
    Does the transition item correlate with the PROM score at baseline?
    4.
    Is the correlation of the transition item with the PROM change score appreciably greater than the correlation of the transition item with the PROM score at follow-up?
  • Data Validation

    Every data point extracted by each reviewer is compared to identify discrepancies using a complex algorithm.
    Once discrepant data points have been identified, reviewers meet with a third senior reviewer to resolve discrepancies.
    When required, a fourth or fifth reviewer may study the primary publication to ensure the accuracy of the data included in PROMID.