Journal of the College of Community Physicians (Dec 2005)

Development and validation of a scale to measure the perceived access to medical care

  • P. R. Wijesinghe,
  • R. de A. Seneviratne,
  • R. L. Jayakody

DOI
https://doi.org/10.4038/jccpsl.v10i1.8267
Journal volume & issue
Vol. 10, no. 1
pp. 18 – 29

Abstract

Read online

IntroductionInstruments measuring access to medical care are available. However. development of new instruments suitable to the specific population and validation of it to that population is essential. The objective of this study was to develop and validate an instrument measuring the access to medical care. MethodsA Likert scale of 35 items was constructed and validated using 178 clients of four private pharmacies and six hospital out door pharmacies in an urban and rural district. Construct validity was assessed using exploratory factor analysis, item analysis and measuring internal consistency. To measure the tendency to agree with statements regardless of their content (Acquiescent Response Sets-ARS), matched pairs of items were included. ResultsFactor analysis extracted four main factors which accounted for 41% of variance. Items of "availability of services" formed the first scale. However, items related to regularity of service provision hypothesised in the "availability" dimension merged with hypothesised items of "acceptability of Services ". This scale was named "Acceptability of Services and Regularity of Service provision (AC & RS)". Third scale which consisted of three items hypothesised to be in the dimension of "Acceptability of services " was logically named "Concern for Clients (CC)" Fourth scale comprised "Affordability" items. One hypothesised item of affordability loaded onto the factor of "Availability of Services". All item-scale correlations were above 0.3. Higher correlation of the majority of items (33) with the tentative scale than with other scales confirmed the factor scales. Factor scales were internally consistent as Cronbach's alpha exceeded the accepted level of 0.5. None of the dimensions proved to be sub scales of higher order scales as inter scale correlation did not meet Anastasi's criteria of 0.5-0.7. Reliability of the ARS scale was low (0.45). ConclusionsExtracted factor structure was consistent with theoretically hypothesized dimensions with a small difference. Two small multi-item dimensions "Affordability" and "Concern for Clients" need further validation with more items. To in­ crease the reliability of ARS scale. more matched pairs are necessary.

Keywords