Reliability and validity of the Delphi method in guideline development for family physicians

Tomasik, Tomasz PhD MD

Lecturer, Family Physician, Department of Internal Medicine and Gerontology, Jagiellonian University Medical College, Krakow, Poland

ETHICS

Ethical Committee approval was not obtained because the study did not involve patients, human material or personal data and was lawfully available.

PEER REVIEW

Not commissioned; externally peer reviewed.

CONFLICTS OF INTEREST

None.

ADDRESS FOR CORRESPONDENCE

Tomasz Tomasik, Department of Internal Medicine and Gerontology, Jagiellonian University Medical College, ul. Bocheńska 4, 31–061 Krakow, Poland. Tel: +48124305593; fax: +48124305584; email: [email protected]

Received 11 November 2009

Accepted 6 August 2010

Quality in Primary Care 18(5):p 317-326, 2010.

Background

The aim of this study was to assess the reliability and validity of the Delphi method in preparing hypertension guidelines for family physicians in Poland.

Methods

A two-round Delphi method was used to develop guidelines for management of hypertension by the College of Family Physicians in Poland. The respondent group consisted of 36 family physicians and 19 specialists. To assess reliability, the internal correlation coefficient (Cronbach’s alpha) was calculated for both rounds. Criterion validity was examined by comparing recommendations in the guidelines with scientific evidence from a study in the Medline database. The content validity was estimated by measurement of extent to which recommendations were related to the competences of family physicians in Poland. To calculate construct validity, a comparison between the results obtained by the Delphi method with those of other methods was performed (Holley and Guilford (G) index).

Results

Reliability (Cronbach’s alpha) was 0.944 for round one and 0.850 for round two. The five missing recommendations and contrary evidence for one recommendation indicated that criterion validity was fulfilled only in part. Among 28 routine tasks related to hypertension carried out by family physicians, the guidelines contained no recommendations for five (17.9%) of them. This indicated that content validity was only satisfactory, whereas construct validity was good (Holley and Guilford (G) index 0.39–0.65).

Conclusion

The Delphi method in the form used in Poland for preparing hypertension guidelines showed good reliability and satisfactory validity. The method should be used judiciously and only after careful preparation.