Potentially Inappropriate Medication Prescribing for Elderly Outpatients in Emilia Romagna, Italy

As defined in the Beers’ criteria, ferrous sulphate should not exceed a dosage of 325 mg/day because constipation is likely to occur in the elderly with a higher dose. Ferrous sulphate is sold in Italy as Ferrograd® (the use of trade names is for product identification purposes only and does not imply endorsement) [Abbott S.p.A], which contains 40 tablets of 525mg (the tablets are not divisible). Although no specific information on the daily dose taken by a patient was available in the database, each tablet of ferrous sulphate itself potentially exceeds Beers’ criterion for this drug. We felt, therefore, that for purpose of this study, ferrous sulphate should be considered as a drug to be avoided in the elderly, especially because safer iron supplements are available through the Italian National Formulary.

The following medications were excluded because they were not reimbursed by the Italian National Formulary in 2001: propoxyphene, oxybutynin, flurazepam, meprobamate, lorazepam, oxazepam, alprazolam, temazepam, triazolam, dipyridamole, reserpine, clinidium-chlordiazepoxide, chlorpheniramine, diphenhydramine, hydroxyzine, cyproheptadine, promethazine, ergot mesyloids, cyclandelate, meperidine (pethidine), laxatives and mineral oil.

In Emilia Romagna, plain locations are more likely to be associated with urban areas, whereas hill or mountain locations are more likely to be associated with rural areas.

A commune is the smallest administrative district.

Our analysis confirmed that the two measures do indeed measure somewhat different independent effects (see Results section and the table for the multivariable logistical regression analysis in that section) and that there is no evidence of multicollinearity. Multicollinearity leads to inflated variance estimates. If multicollinearity were a problem, we would have seen unusually large coefficient standard errors and removing one of the two variables would have lowered the size of the remaining standard errors. This did not happen in our regressions. Moreover, the fact that in our most comprehensive logistical regression model the CCDG score was still statistically significant after controlling for the number of medications suggests that the individual’s severity of disease — at least as measured by the CCDG score — was associated with inappropriate medication prescribing.