In Barcelona, AIDS has been a mandatory notified disease since 1986, and HIV was included in the registry as a voluntary notifiable disease only until 1996 . During the HIV/AIDS epidemic there was, and continues to be, controversy over whether HIV infection should be notified. Following a long debate over many years, in 2010 Catalonia passed a law making a mandatory registry of HIV infection . Such a mandatory registry permits better characterisation of the infection, avoidance of duplication, the study of sexual partners, effectiveness of therapy, and follow-up of cases of infection until the appearance of AIDS, or death . However, some studies have suggested that the introduction of this type of notification could diminish the number of persons who undergo the test; however these predictions have not been reflected in significant declines in testing in the United States .
In our study, both the results of bivariate analysis and of time series analysis showed that the introduction of mandatory notification of new diagnosis of reported HIV infection led to an increase in such notification of almost three times. This increase in the completeness of notification results in a more precise description of the epidemiological situation regarding HIV infection. The epidemiological characteristics are similar in both periods, although some proportions have changed, new diagnoses being more common among men with higher risk of sexual transmission; particularly MSM and those aged over 30 years. There is a notable percentage increase of groups at risk following the introduction of the law, such as individuals aged under 30 and people born outside Spain. Regarding the notifying centre, hospitals are the main sources of notifications in both periods.
Comparison of the epidemiological characteristics against data available at national level (2003–2010) revealed differences with respect to the proportions in the different categories of transmission. It is important to stress that national-level data cover 71% of the population (32,843,416 inhabitants in 2010) , making it difficult to extrapolate to the rest of the country, since the degree and patterns of HIV transmission differ between the different Autonomous Communities. It should also be noted that amongst them, the notification of new cases of HIV is voluntary and that some proportion of these cases have probably not been notified . This reveals the need to have a notification system that permits more exhaustive surveillance of the epidemic, in order to obtain appropriate quality and quantity of information. Thus, this information will allow the estimation of the magnitude of the infection and to monitor its trends , especially in the large cities, which is where cases tend to be concentrated. Comparing with the data of HIV diagnoses that were reported by the EU/EEA countries in 2011 (rate of 5.7 per 100000 in the population), 10% of all HIV diagnoses were reported among young people aged 15 to 24 years, and 30% among MSM . These trends are similar to those observed in Barcelona, describing a concentrated epidemic especially by sexually transmitted infection, mainly in MSM.
Studies with regard to the switch from voluntary to mandatory notification were not possible to find. However, in other settings, it has been observed that appropriate and exhaustive systems of surveillance of transmissible infections, not only HIV, generate more precise information at population level, presenting a more realistic picture of the infection situation [22, 23]. Accurate information has helped large scale efforts in eradicating smallpox  and reducing dracunculiasis in Asia and Africa [25, 26].
Reporting of HIV infection must be interpreted with caution however, taking into account other available epidemiological data, because these reports do not provide a direct measurement of the incidence or prevalence of HIV infection [12, 13]. The proportion of HIV infected individuals, who are diagnosed and reported, varies according to the phase of the epidemic , HIV testing patterns , and characteristics of surveillance systems. Therefore, new diagnoses of HIV infection, reported into a registry, varies despite of their mandatory or voluntary state. But with a mandatory notification registry, all cases that have been diagnosed are reported, despite their infection state. HIV mandatory reporting is helping to improve assessment of the scale and extent of recent HIV transmission in the population , to visualize a better picture of HIV endemic situation in the city and to describe the epidemiological characteristics of emerging risk groups. It is important to stress that with mandatory notification, elimination of duplicate reports and the ability to match reports of HIV infection with other data sets are accomplished . These are two characteristics , which are essential requirements, for an effective HIV reporting system. Another important advantage of mandatory notification is that it is possible to recover data by active surveillance; therefore there might be a decrease in missing data as it was observed in our analysis.
Strengths and limitations
Time series analysis has permitted the description and estimation of the evolution in trends affecting the notification of new diagnoses of HIV infection [29, 30]. Thus, while on the one hand we observe an increase in the number of mandatory notifications, the slope of the overall trend persists in both periods.
Among the limitations of this study, one is the short period of follow-up of mandatory notification. This raises the issue of the need to continue its evaluation, in order to determine whether the changes and the rise in number of notifications persist over time.
Another aspect to be taken into account is that the analysis was conducted with the start date of mandatory notification as January 2010, despite the fact that the law only came into force in July 2010, and the publication of the 2010 manual of notifiable diseases and mandatory notification of HIV infection was also included . Consequently, some centres, not notified when this was voluntary, began to do so at the beginning of 2010. We decided that the first two quarters of 2010 were included in the analysis, as the impact of mandatory notification starts from the second quarter of this year, despite the law only coming into force in July 2010.