Patients and Methods
This study was conducted in a 60-bed geriatric unit of a university hospital (mean hospitalisation duration about 11 days), which is a 2362-bed tertiary hospital. All patients over 70 years of age admitted consecutively from November 1997 to December 1999 were included in the study. All study participants were admitted directly through the emergency unit of the hospital. Only ADEs present at admission were considered for this study.
Data were collected and analysed by two physicians and one pharmacist twice, first on admission and again when the patient was discharged. One physician and the pharmacist recorded data (drugs consumed, signs or symptoms at admission) during the two days following admission. At the end of hospitalisation, a second physician and the pharmacist systematically checked the list of drugs initially recorded for each patient, to complete the list if necessary. They also checked whether the signs that were initially attributed to an ADE had disappeared. The collection of data by two physicians was especially necessary in cases of neuropsychological disorders or cognitive impairment.
All drugs administered during the month preceding hospitalisation were recorded. Information about treatment was obtained from the patient and the patient's general practitioner. Additional information was obtained from the family when the medical prescription was not available or when the patient had cognitive or vigilance impairment. The investigation only took into account drugs that were consumed and not drugs that were prescribed but not consumed. Voluntary drug intoxication was not taken into account in our study.
Reference data were taken from the Vidal dictionary, which corresponds to the Physician's Desk Reference, and is the reference book used for drug prescriptions by French physicians; it is updated annually by the French Health Department.
When signs or symptoms suggested an ADE attributable to administered drugs, the investigators determined the responsible drug or drug combination by using the algorithm of Bergman and Wiholm.[18,19,20] Signs and symptoms were classified into 10 groups of disorders: metabolic and/or renal (dehydration, ionic disorders, renal failure); cardiovascular (orthostatic arterial hypotension, heart rhythm or conduction impairment); neuropsychological (vigilance impairment, delirium, agitation); digestive (diarrhoea, vomiting); haemorrhagic (digestive or cerebral bleeding); cutaneous (rash, pruritus, urticaria); urinary (urinary retention); pulmonary (respiratory failure); or haematological (agranulocytosis, pancytopenia).
Initially only 'definite' and 'probable' ADEs were taken into account but, due to the lack of rechallenge, only signs or symptoms classified as 'probable' ADEs were finally taken into account. A sign or symptom caused by one drug or a drug combination was considered a 'probable' adverse effect if: (1) there was a reasonable temporal sequence from the commencement of the drug combination treatment, (2) there was a known response pattern, (3) if the signs or symptoms were improved by discontinuation of responsible drugs, (4) the signs and symptoms could not reasonably be explained by the known characteristics of the patient's clinical condition.[19,20]
To evaluate the involvement of drug combinations, all possible two-by-two combinations were considered for each patient. When the drugs of a drug combination had been prescribed successively, and not simultaneously, we only considered signs or symptoms occurring during the time of administration of both drugs.
Severity of ADEs was classified into two groups: severe ADEs and moderate ADEs. An ADE was severe when signs and symptoms were directly responsible for hospitalisation and/or life threatening or fatal. An ADE was moderate when it was associated with another disease and was not directly responsible for hospitalisation.With regard to the evaluation of risk factors, we only took into account those risk factors related to drug prescriptions or patients' conditions: excess doses of drugs based on reference data, potential DDIs, acute intercurrent diseases (e.g. infection, dehydration) or interfering chronic diseases not taken into account by the physicians (e.g. obstructive bronchial disease and -blockers, prostate adenoma and anticholinergic drugs). In this latter case, the physicians were contacted by the investigators to confirm that they had prescribed the drug without considering the interfering chronic disease.
We did not evaluate poor compliance. This factor occurs frequently and has been reported in about 60% of the elderly.[3,4,9,21] The main risk factors for poor compliance in the elderly are cognitive impairment, treatment with psychotropic drugs or a lack of patient information.[3,4,9,21] However, patients generally do not want to admit to poor compliance or do not remember it.
Some risk factors can be considered preventable, for example, excess doses of a drug (according to the Vidal dictionary, which takes into account the dosage adaptation according to age and/or renal function); interfering chronic diseases that were not taken into account; and preventable DDIs. Preventable DDIs were defined based on absolute or relative contraindicated drug combinations [e.g combination of an angiotensin-converting enzyme (ACE) inhibitor with potassium, or the combination of a nonsteroidal anti-inflammatory drug with an antihyperglycaemic sulphonylurea].
Other risk factors that were considered to be non-preventable were acute intercurrent diseases (e.g. infection, dehydration) and non-preventable DDIs. Non-preventable DDIs were defined as resulting from synergic drug combinations or drug combinations adapted to the clinical state that only required precautionary use (e.g. the combination of a digoxin preparation with diuretics, or the combination of a benzodiazepine with an antidepressant agent).
Patients might have both preventable and nonpreventable risk factors: for example, acute renal failure resulting from acute dehydration caused by infection and an inappropriate drug combination of an ACE inhibitor and a nonsteroidal anti-inflammatory drug. Thus, combined risk factors were considered preventable when all risk factors were preventable.
Data were recorded and analysed using the Statview® statistical package. Qualitative values were compared using the Chi-square test. A p-value below 0.05 was considered statistically significant.
Clin Drug Invest. 2002;22(6) © 2002 Adis Data Information BV
Cite this: Preventable and Non-Preventable Risk Factors for Adverse Drug Events Related to Hospital Admission in the Elderly - Medscape - Jun 01, 2002.