Review of the use of resource use instruments based on patient recall in relation to other methods of resource use estimation

Background

We conducted a review of articles citing trial-related resource use measures catalogued in the Database of Instruments for Resource Use Measurement (http://www.DIRUM.org). The aims were to assess: (i) how the instruments were used in practice; (ii) which items of resource use were most frequently measured using patient-recall; and (iii) how estimates compared if more than one method of data collection was used for the same resource items

Methods

Articles citing references to measures were identified using Google Scholar, ISI Web of Science and Scopus. These were reviewed and screened according to inclusion criteria. Data were extracted on: the method of administration, resources measured, rates of return and the nature of the other methods of resource use measurement.

Results

Nearly all citing articles (143/146) used an instrument derived from Beecham and Knapp’s Client Service Receipt Inventory or a variation thereof (81/143). Most measures relied on patient- or proxy-recall (126/146) generally administered during researcher interviews. Use of primary and secondary care was the most widely reported items of resource use (136/146). Twelve studies compared more than one method of data collection for the same resource items and 8/12 showed good correlation between medical records and patient/carer recall or at least indicated one could routinely validate the other. No single administration method was deemed superior.

Conclusion

Patient interviews are the most common method of questionnaire administration but studies lack clarity on how they are conducted. Results indicate good precision and accuracy in recall questionnaire use; however, concerns about recall bias still exist.