Study for Early Detection of Drug Interactions in Older Hospitalized Patients Using on Line Software

Transversal, Multicentric Study for Early Detection of Drug Interactions in Older Hospitalized Patients Using on Line Software in Córdoba, Argentina

A drug interaction (DI) is the mutual action of two drugs in a way that they can increase their action, even to a toxic level, or reduce it to its minimum.

People elder than 65 years old have theirs biological ability to metabolized and eliminate drugs impaired. Even more, they tend to suffer from many diseases, are treated for many physicians, and receive many drugs for those conditions. If hospitalized older people are prone to receive a greater number of drugs. This scenario is the worst to suffer from adverse drug events and DI, which in turn compromise more the health and even life of hospitalized older people.

Many computerized strategies have been developed to prevent those problems. In this trial the investigators use on line software to early detect DI that could endanger health or life of hospitalized older patients.

A drug interaction (DI) is the mutual action of two drugs in a way that they can increase their action, even to a toxic level, o reduce it to its minimum.

Adverse drug events (ADE) are an important health problem. In the book To Err Is Human: Building a Safer Health System its authors Kohn, Corrigan and Donaldson from the "Committee on Quality of Health Care in America, Institute of Medicine" (U.S.A.) expressed that 2 out of every 100 hospital admissions were due to ADE. That results in an average increase of U$D 4,700 extra for each admission or about to 2.8 million U$D annually for a 700-bed teaching hospital. In this study were only considered direct costs. No indirect costs were taken into account, like days out of work, rehabilitation, deaths, lost of quality of life, etc.

In 2003, in a cohort study done in people of Medicare elder than 65 years old, were detected 50.1 ADE / 1000 persons a year. Of that number of ADE 13.8 were considered preventable. In Switzerland 3.3 % of hospitalisations were considered due to ADE.

Among ADE, DI are a very important part of the problem. In Canada was demonstrated that certain drug associations were responsible of an increase in the number of hospital admissions. They would not have occurred if those prescriptions would have been properly controlled or never done. In Mexico 3.8 % of patients elder than 50 years old received prescriptions of drugs, which should be avoided according to their interaction. In hospital setting DI endanger even more health and people's life. In Spain was demonstrated a prevalence of clinical relevant DI of 3 % in hospitalized patients, and in Switzerland that percentage rose 11 %. This phenomenon increased in later years, in Rotterdam DI compromising lethal risk in patients elder than 70 years old, climb from 1.5 % in 1992 to 2.9 % in 2005. Drug interactions are directly proportional to: the number of administered drugs (> 2 drugs or > 4 ); the age of patients and the number of prescribing physicians.

Older persons, for its detriment in physical condition and the pathologies they suffer from, are prone to suffer DI when being prescribed with drugs or even herbal medicines. In Brazil ambulatory patients elder than 60 years old have 2.5 pathologies diagnosed and consume in average 1.3 to 2.3 drugs. During hospitalisation the number of prescribed drugs reach 9.9 to 13.6 per patient. In a survey done in six European countries the number of drugs consumed for each ambulatory patient was 7. In Argentina according to the Instituto Nacional de Estadísticas y Censos (INDEC) the number of citizens elder than 65 years old was 3,587,620 (9.89% total population). Unfortunately there is no other information about polypharmacy or prevalent pathologies in INDEC or other medical databases or national institutions web sites.

In 1995, it was published an article in JAMA that stated the possibility that 28 % of ADE could be prevented in non-obstetric hospitalized adults. Fifty six percent of those were due to mistakes done during prescription, 34 % during administration, 6 % during transcription of indications and 4 % during dispensation. This astonishing results poses on the need to develop strategies to cut this numbers.

Many lists of drugs with potential to produce ADE were developed to prevent DI and other related drugs damages. Beers criteria (BC) is one example of them. In Amsterdam it was demonstrated an 80 % improvement in drug regimens in patients elder than 81 years old of in whom their therapeutic indications were reviewed and adjusted using BC and the Medication Appropriateness Index after suffering from ADE. In that study, a list of only 10 drug combinations was also used but did not showed the same benefits. In Argentina, the Adminstración Nacional de Medicamentos, Alimentos y Tecnología (ANMAT) publish in its web page a list of 16 risky drugs, considering the disadvantages of their own pharmacological action, but there is no information about its usefulness.

Much software was developed in later years to alert physician about DI. Some of them are even more sensible than the search of DI at the bedside, but this software overestimates the prevalence of serious DI. The usage of computerized alert systems for prescriptions significantly reduced potential ADE, however many physicians tend to override them for different reasons. Finally a Cochrane´s review showed a reduction in hospitalization days and drug toxic effects when using a computerized warning systems for drug dosing.

The growing number of drugs, apart from its not always well-proven benefits, implies risks for health; specially in sick and older people and even more during hospitalizations. Nowadays software to alert physicians about DI or ADE could help to prevent them but its real contribution is still a matter of debate. For this reasons we developed this trial to test the usefulness of on line software to detect DI in hospitalized elder patients.

Inclusion Criteria:

• Hospitalization
• 65 years old or more

Exclusion Criteria:

• None