Psychometric Analysis of an Advance Directive

Hillel R. Alpert*, Herbert Hoijtink, Gary S. Fischer, Linda Emanuel

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

13 Scopus citations


OBJECTIVES: Reliability and validity are as necessary for predrafted advance directive forms as they are for all clinical assessment instruments. Performance of predrafted advance directive forms with both lay persons and clinicians is relevant. Evidence relating to test-retest reliability, content validity, and criterion-related validity of one form, the Medical Directive, has been documented for outpatients. The authors investigated construct validity and external validity among outpatients, physicians, and the general public. METHODS: Four hundred ninety-five outpatients, 513 physicians, and 102 members of the general public were surveyed with the Medical Directive. Preferences for 11 specific treatments in four to six illness scenarios were recorded. Mokken modeling of responses was used to produce a psychometric scale of receptiveness-to-treatment and desirability of treatments. The Kuder Richardson-20 statistic, Friedman's procedure for analysis of variance, and the Kruskall-Wallis test were used, respectively, to measure inter-item reliability, the relation with scenarios, and the relation between physicians' general goals for care and their scaled preferences. RESULTS: All model diagnostic tests indicated a close-fitting scale for all three respondent groups. Kuder Richardson-20 for outpatients (.98), physicians (.97), and the public (.93) demonstrated high inter-item reliability. Treatment desirabilities were related to invasiveness. Receptiveness-to-treatment was related to prognoses and disabilities of described illness scenarios among each group and to physicians' goals for care. CONCLUSIONS: The Medical Directive has construct validity in relations among specific treatment preferences and between treatment preferences, illness scenarios, and goals for care. External validity is supported by study of separate outpatient, physician, and general public populations. The treatment items constitute a highly reliable scale that can be used in further empirical research regarding life-sustaining treatment.

Original languageEnglish (US)
Pages (from-to)1057-1065
Number of pages9
JournalMedical care
Issue number10
StatePublished - Oct 1996


  • Advance directives
  • Attitude
  • Ethics, medical
  • Psychometrics

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health


Dive into the research topics of 'Psychometric Analysis of an Advance Directive'. Together they form a unique fingerprint.

Cite this