The conclusion that is drawn up from the results of the bibliographic review is that there is no one ideal instrument for implementing post-introduction observation. The choice of one instrument or another depends, in great part, on the information to be obtained, the technology sought to be observed, the individual institution’s/organisation’s/centre’s structure/approach, as well as the structural, organisational and financial means available for undertaking such monitoring.

Before starting a post-introduction observation, the information sought to be collected, existing possibilities and available means must be comprehensively assessed, and the design of the data-collection system properly planned.

It is generally accepted that post-introduction observation should be based on active data-collection because existing evidence on passive systems, including surveillance systems, suggests that these furnish more information than do voluntary systems [28,29].Passive systems suffer from an important degree of under-reporting world-wide and may be biased by incorrect information (15). En una nota informativa publicada recientemente por el "British Medical Journal" se estima que sólo se notifican entre el 5 y el 10% de los efectos adversos graves (16). The reason for this lack of reporting has been well documented in the case of pharmaceutical research and could clearly be applicable to other health technologies (30).Among other things, reasons for under-reporting include complacency, fear of medical-legal implications, and uncertainty as to whether the medication or device has really been the cause of the adverse effect.

Different experiences highlight the fact that the experience and collaboration of clinicians, technical staff, patients and health decision-makers is an absolute requirement for achieving success in any safety-related initiative. An observation system will only be successful if it is implemented by taking into account the needs and experiences of clinicians and patients (28).

Bearing in mind prior experiences in monitored use (6) and the results of the bibliographic review, we put forward that, in our context, post-introduction observation could be undertaken by means of clinical registries, using questionnaires completed by clinicians for collecting data of an administrative and clinical nature at the time of short-term intervention/treatment and telephone surveys of patients for medium/long-term follow-up. Telephone surveys are also proposed as the tool of choice for obtaining information on the technique’s diffusion.

We recommend that the questionnaire should be in paper rather than electronic format, since the conclusion drawn from the bibliographic review is that the latter seems to register very low response rates, a result largely confirmed in monitorised use.

Based on the fact that there is evidence to show that the degree of reporting is higher when information is obtained directly from the patient, and taking into account the fact that monitored use experiences identified important problems of continuity, we suggest that medium-/long-term information be obtained directly from patients (6). Although the evidence indicates that face-to-face interviews yield higher response rates, the geographic dispersion of the population makes this option enviable due to the considerable time and high cost involved. Bearing in mind that the greater part of the population has telephone access and that the non-response rate is lower than that of postal surveys, we view this as the most suitable option. Nevertheless, it is necessary to recall here that the information drawn from surveys is context-specific. It largely depends on the prior information received by the patient, his/her motivation, and the interviewer’s skills. Accordingly, this calls for collaboration from the clinician to increase the response rate, and for interviewers to have adequate training in conducting personal interviews.

In future, electronic medical records could be the tool of choice for post-introduction observation. At present, however, in view of the fact that in many countries, including ours, such histories are still in the implementation stage and that in others they are still in the initial stages of development, it is too soon to for this data-collection instrument to be considered.