ADVERSE EVENTS: RELATED TO DRUGS OR IATROGENIC?

Abstract

To the Editor: Dr. Gerety et al published an interesting study, “Adverse Events Related to Drugs and Drug Withdrawal in Nursing Home Residents,”1 investigating the epidemiology of untoward drug effects with respect to the severity of these events and to the associated risk factors. The study design is very similar to one we published previously in the Journal, “Adverse Clinical Events in Long-Term Nursing Home Residents.”2 We would like to make some comments about Dr. Gerety's study regarding the methods and the underlying philosophy. First, severity of adverse drug-related events was dichotomized into major and minor along the dimensions (1) physician visits, (2) hospitalization, and (3) prolongation of hospitalization (as judged by the investigators). The authors do not offer any detailed criteria that elucidate these dimensions. Thus, the attribution of an event to one of the two categories could be aleatory as such a rating method is influenced by occasional factors and circumstances extrinsic to the event itself. Moreover, the above mentioned dimensions do not measure any clinical and functional aspects. Thus it is impossible to identify any reliable cause-effect relationship between drug exposition/withdrawal, onset, and outcomes of the adverse event. This study was conducted without any face-to-face patient assessment. Furthermore, there was no formal evaluation of reliability regarding classification and determination of the probability of the relationship between drug use and adverse events. These are all important methodological biases that severely limit the internal validity of the study. Secondly, we are extremely surprised by the high incidence and severity of the adverse drug events. In our study we managed 273 adverse clinical events in 94 chronic patients during a 3-month period. Only three (1%) of these events were considered iatrogenic (all drug related), an incidence of 0.011 patients/month. Stringent nurse-physician monitoring for more than 7 days was required in one event only, related to an overdose of phenothiazines. These data are very different from those reported by Dr. Gerety. The same discrepancy emerges from another study that has just been completed in our setting on hip fracture patients referred for rehabilitation in a post-acute stage. Eighty-nine patients experienced 427 adverse clinical events overall (2.1 patients/month incidence). There were 15 iatrogenic events related to drugs (3.5%), ie, incidence of 0.07 patients/month. No iatrogenic event caused new functional impairment. NSAIDs, L-Dopa, antidepressants, and digitalis were the drugs most frequently involved. Thirdly, in our opinion, adverse events related to drugs should be considered more appropriately within the context of iatrogenesis (defined as any untoward effect induced by any medical treatment). Iatrogenesis is the “dark side” of therapy: the effect (therapeutic or adverse) of a medication cannot be completely understood as long as it is evaluated separately from the illness that it has been prescribed for. Iatrogenesis in comprehensive geriatric assessment constitutes one dimension only of the complex domain of clinical instability. Therefore, assessment of untoward drug effects cannot be isolated from the management and monitoring of intercurrent illnesses. If it is, it cannot be considered a useful tool for drug management in the elderly. Clinical pharmacology needs indicators on how, how intensively, and why a drug acts with respect to severity of the illness, the burden of comorbidity, other therapeutic procedures, and outcomes. The method proposed by Gerety et al for the evaluation of adverse drug events does not allow any of these variables to be controlled; thus, it cannot adequately address these crucial problems in geriatrics. Fourth, we believe, as does Brennan,3, 4 that iatrogenesis is closely associated with substandard care in the hospital as well as in the nursing home. Accurate assessment and continuous patient evaluation help to prevent adverse clinical events and to limit their severity. This is the only strategy that will limit the burden of iatrogenesis in geriatrics. Therefore, we do not agree with Gerety, who states that risk factors such as comorbidity may not be amenable to intervention. The improvement of care management in chronic patients is feasible only if the geriatrician is directly and fully engaged in the decision-making of medical and nursing care. As a consequence, the prescribing and utilization of medication in nursing homes will automatically improve. Editors note: The above letter was referred to the author of the original paper, and Dr. Gerety's reply follows.