PRO, CRO, and Laboratory tests / device measurements

In an article by Willke et al (Controlled Clinical Trials, 25, 2004), the study endpoints were classified as three major categories. Endpoints were classified into the following three major categories, and the presence or absence of each of these categories was noted for each product reviewed. Each product may have employed one, two, or all three types of endpoints:

• Laboratory tests and device measurements,
• Clinician-reported outcomes (CROs)
• Patient-reported outcomes (PROs).
  

Laboratory and device measurements included highly objective typically numerical measures often performed by machine.

Clinician-reported outcomes included those that might be considered traditional endpoints, either observed by the physician (e.g., cure of infection and absence of lesions) or requiring interpretation by the physician (e.g., radiologic results and tumor response). In addition, CROs included both formal and informal scales completed by the physician using information about the patient. CROs requiring patient input are distinguished from clinicianadministered PROs in that the former requires clinician judgment or interpretation when recording answers, while the latter involves recording precise, unmodified patient responses to prespecified questions.

Finally, endpoints classified as patient-reported outcomes included formal health-related
quality of life measures and any other endpoint that was primarily based on a direct patient report. PROs categorized as "formal" scales are those multiitem questionnaires that have a well-defined standardized format, well-documented procedures for administration and scoring, demonstrated reliability and validity, and some guidelines for interpretation of scores. Other PROs included informal symptom scales, patient global assessments, or visual analog scales, as well as patientreported endpoints recorded in event logs (e.g., specific events). In some cases, nonclinician proxies reported the outcome from the perspective of the patient (e.g., when vaccines were tested in infants); these endpoints were considered patient-reported.