A Clinimetric Approach for Improving the Measurement of Pharmacophobia with Replication in Two Other Samples

stand how we have used clinimetric principles to refine the definition of pharmacophobia, it is necessary to detail the problems involved in its standard use. The original DAI-10 was developed in such a way that it was not possible to hold both positive and negative attitudes toward psychiatric medication simultaneously [4]. A patient’s attitude toward psychiatric medication was defined along a unidimensional continuum that can range from extremely negative (–10) to extremely positive (+10), but it would not be possible for the same patient to hold both positive and negative evaluations of medication at the same time. One might consider the midpoint of this unidimensional continuum of psychiatric drug evaluation to indicate the concurrence of positive and negative attitudes, but mid-range responses may also indicate that there is no opinion about or an attitude of indifference towards psychiatric medication. Therefore, the unidimensional continuum does not allow differentiating those patients who have positive and negative attitudes towards psychiatric treatment from those who are unconcerned or indecisive about it. The aim of this research in the 3 previously studied samples [8] is to improve the DAI-10 scoring system by using a clinimetric plan including 6 successive steps: (1) eliminating 2 confusing items from the DAI-10, (2) performing an exploratory factor analysis of the 8 remaining items to rate 2 factors (liking and disliking medications) in the 3 sample groups, (3) testing based on improved internal consistency of the 2 new subscales versus the DAI-10 in the 3 sample groups, (4) identifying clinimetric pharmacophobic patients (high in disliking and low in liking medications), (5) identifying patients with consistently poor adherence, and (6) testing within the 3 samples for more accurate prediction of nonadherence using the new clinimetric definition of pharmacophobia versus the old definition of pharmacophobia. This cross-sectional, cross-cultural pharmacopsychology study included a total of 1,320 psychiatric outpatients recruited on the Canary Islands (Spain), Mendoza (Argentina), and Mérida (Venezuela) and has been described previously [8]. Two scales were used: the Spanish validated version of the DAI-10 and the Spanish version of the validated Sidorkiewicz instrument [6] to assess treatment adherence for each individual drug taken by a patient. The online supplementary Methods and Materials provide details regarding study design, participants, scales, and statistics with a carefully detailed description of the clinimetric plan. The online supplementary Results include a descriptive section summarized in online supplementary Table S1; a two-factor analysis solution (second step of the clinimetric plan) summarized in online supplementary Table S2; the improved internal consistency of the 2 new subscales (third step) summarized in online supplementary Table S3; the new clinimetric definitions, compared with Treatment adherence [1] may be an important topic in clinical pharmacopsychology [2] (see online suppl. Introduction; see www.karger.com/doi/10.1159/000495940 for all online suppl. material). The Drug Attitude Inventory (DAI-10) [3] is a widely used self-report questionnaire with true/false statements about the perceived effects and benefits of psychiatric drugs, with which patients can agree or disagree. It includes 6 positively phrased items, referring to symptom reduction, and 4 negatively phrased items, referring to side effects [3]. A score is obtained by summing the positive and negative responses, which could range from –10 (very poor attitude) to +10 (best possible attitude), thus classifying patients as “pharmacophobic” or “pharmacophilic” according to their negative or positive score. Pharmacophobia is one of the major variables associated with poor adherence [4] when measured by self-report in our Spanish studies, within the context of the Health Belief Model [1]. In a multivariate analysis of 940 psychiatric outpatients from Spain [5], we found that pharmacophobia was associated with poor adherence; the adjusted odds ratio (OR) was 2.54 (95% confidence intervals 1.84–3.50) after adjusting for the other 3 psychological dimensions known to influence poor adherence. As with other researchers [6], we believe that, despite the DAI’s great usefulness, there is room for improvement even without modifying its original questions, but rather, by improving and simplifying the scoring procedures. Pharmacophobia is what Feinstein [7] would call a clinimetric general index since it is a measure of general health and functional states that are not distinctive for a particular disease or condition but influence adherence to all kinds of medications, not just psychiatric medication. To underReceived: March 4, 2018 Accepted after revision: November 21, 2018 Published online: January 16, 2019