Hospital pharmacists have long identified the relationship between antibiotic consumption and reduced bacterial sensitivity based on the results of the antibiotics. Over the past decade, clinicians and researchers have focused on studying the most effective causes and ways to prevent and eliminate antibiotic resistance. Population monitoring is a relatively new method and its implementation has become practically feasible with the introduction of electronic databases. This article provides an overview of the data collection and analysis methods used in this method.

Accounting for antibiotic purchases is the easiest way to measure consumption of antibiotics. However, this is an indirect method which largely depends on the stability of the institution's purchases, producer prices and supplier discounts. This estimate is not stable enough to compare the consumption of antibiotics within or between institutions.

To estimate consumption, a sum of the number of grams of antibiotics administered or prescribed can be used. This method has an advantage over supply in that it does not depend on the price of the drug, but it still does not allow to compare the consumption of antibiotics of different activities.

The most widely accepted method for evaluating antibiotic consumption is DDD (defined daily intake). This method was adopted by WHO more than 20 years ago as an option for standardizing studies on the use of drugs in various countries. DDD theoretically corresponds to the average maintenance dose of the drug when used according to the main indication. The World Health Organization has established and regularly reviews DDD for the major antibiotics used in the United States. The unit traditionally used is the number of DDD / 1000 patients / day.

The DDD / 1000 patient / day score has its limits. Thus, the DDD value is based on the usual (standard) dosage for an adult with normal kidney function. In facilities with a high proportion of children or patients requiring hemodialysis, DDD may not accurately reflect the use of antibiotics.

There are other methods than DDD: taking into account the days of treatment, prescribed daily doses, average daily doses. When comparing different studies or levels of antibiotic consumption, it is necessary to consider the evaluation method used.

The simplest way to assess the resistance of microorganisms is with an antibiotic. When studying the relationship between antibiotic consumption and resistance, as measured by an antibiotic or another method, several aspects must be taken into account.

First, a study in individual units of a health facility, for example in the intensive care unit and the intensive care unit (ICU), in comparison with a complete health facility, allows for more precise identification of the correlation between antibiotic consumption and resistance. A similar study may also reveal relationships that were not found at the level of the entire health facility. This is not surprising, since the highest level of antibiotic consumption and resistance is observed in ICU. Based on this data, it is proposed to monitor antibiotic consumption and sensitivity both individually and collectively at the level of individual health facilities. In addition to evaluating the consumption of antibiotics in a health facility, a comparison can be made of the indicators of the department of this hospital with external data from other similar departments, for which multicenter databases are used, and we check whether the indicators differ.

Second, although the results of antibiotics are easy to obtain, these data cannot reveal small changes in resistance, which are manifested by an increase in the minimum inhibitory concentration or detected by molecular methods.

Theoretically, early detection of resistance problems associated with the consumption of antibiotics allows you to intervene in a timely manner and reduce the incidence and mortality caused by resistance. In the final analysis, it is important to recognize that all laboratory methods, although they do determine antibiotic resistance, are indirect methods for assessing nosocomial infections that are resistant to antibiotics.

Biological model developed in the 1970s. Showed a clear relationship between the consumption of antibiotics and the selection of resistant E. coli in humans. Several models have been found between the effect of the drug and the sensitivity of the microorganism. Resistance also depends on a number of environmental and patient-related factors. Continuous surveillance of antibiotic consumption surveillance is a relatively new industry that has emerged in the past 20 years. Most research models, such as the case-control with regression modeling or simple linear regression, come from traditional epidemiology. Studies show that addiction is more complex than just pushing drugs.

The unresolved question is which studies are best: studying the process in individual patients or a population. In fact, the results of the two types of studies are complementary.

In a case-control study, it is possible to use two different control groups: patients with sensitive forms of pathogens (traditional control) and patients selected in the same rooms as patients in the experimental group, without study of culture. The second type of control is recognized as more relevant.

Simple linear regression can be used to reflect the relationship between antibiotic consumption and resistance, particularly in multicenter studies. Many studies have shown a pronounced correlation between resistance to a particular antibiotic and its consumption, especially when studying data for a particular department. Multivariate analysis is preferable because it reveals the influence of factors other than antibiotics on resistance.

The limitation of case-control studies and studies with regression analysis is the inability to take time into account. The consumption of antibiotics, antibiograms and nosocomial infections vary over time. There is a gap between the consumption of antibiotics and the development of resistance.

Time series analysis is a multifactorial technique that uses mathematical modeling of data series over time. This model can be used to predict future trends of an indicator based on its past changes. The autoregressive integrated moving average model is free from the disadvantages of traditional methods, but their limitation is the need for a significant period of time (usually months) to build the model.

The results of the multicentre NNIS-ICARE study showed that systematic surveillance, comparison with internal and external reference data, the introduction of antibiotic policies lead to a decrease in nosocomial resistance.

There is no optimal method for monitoring antibiotic use and recommendations for choosing an antibiotic whose application needs to be changed. Further research in this direction is necessary.