In the present study DIDs were lower than that previously recorded for Greece . This could be explained in part by the fact that in the previous decade family practice in rural areas has been provided by uncertified physicians. Further in urban areas there isn't an established net of well-trained primary health physicians. Over the last decade family practice has reached a revolution and general practitioners provide primary health care services in rural areas as in our study. Our findings further confirmed the seasonal fluctuation of prescriptions, the increased use of broad spectrum antibiotics and the predominance of prescriptions for respiratory tract infections (82.8% and the 71.4% in children and adults, respectively). This is in accordance with Goosens et al study that included 32 countries . In adult respiratory infections, penicillins, cephalosporins and macrolides-lincozamides shared almost equal amounts of the total DDDs. In contrast, in children penicillins were more commonly prescribed. Macrolides-lincozamides accounted for 48% for respiratory infections in adults and 31.9% in children, considerably higher than the percentages (18%) reported previously for Greece .
The overall DID rate was 23.4. This DID is one of the highest among European Countries, but it should be stressed that it is lower than the proportions given by the ESAC Project Group for Greece . Recently, rural areas in Greece are increasingly being served by physicians specialized in general practice and this may have contributed to more rational antibiotics prescribing.
The assumption although of DDD and DID in children has several limitation problems as DDDs are normally assigned based on use in adults. The Guidelines for ATC classification and DDD assignment  state that the medication dose recommendations will be different depending on the age and body weight of the patient. Because of this, the WHO International Working Group for Drug Statistics Methodology has determined that it is impossible to assign pediatric DDDs, and studies that attempt to investigate drug utilization in children cannot use them. They further recommend that the general DDD be used as a measuring instrument for overall comparisons when the pediatric subgroup is difficult to identify. Therefore in our study the expression of the prescribing DDD in children might have as a result a misserpretation as it relates to an average adult weight. It can be concluded that medication use is considered greater than that estimated, as we cannot estimate precisely DDD and DID in children (DDD is the assumed average maintenance dose per day for a drug in adults). As our results showed no significant differences between adults and children it could be estimated that in children either there was a higher morbidity or there was an overuse of antibiotics. It probably means that much bigger fractions of children compared to adults were treated with antibiotics. Even although with these limitations on DDD and DID use, an overall estimation of antibiotic prescribing can be concluded.
The use of penicillins in children was satisfactory (49.5% of DIDs) but not in adults (31.7%). The overall use of penicillins in our study (38.1 of DDD) was higher than what was previously reported for Greece (31.1% and 21%) by Tzimis et al  and Molstad et al  respectively. In accordance with previous studies [3–5], the use of phenoxymethylpenicillin accounted for 1.5% of the total DDD of penicillins in children and 2.7% in adults. The increased use of wide-spectrum penicillins and particularly of amoxicillin-clavulanate has been also observed in the adjacent rural Anogia Health Centre area . The predominance of broad spectrum agents has been noted in many countries and has been attributed to marketing campaigns, making physicians less sensitive to cost and quality of the prescribed drugs, and the belief that respiratory infections are an indication for antimicrobials [22–24]. All prescribed cephalosporins were of the second generation. No first-generation cephalosporins were prescribed, again a finding suggesting the unjustified overuse of broad-spectrum agents even in simple infections. No third generation cephalosporins were prescribed, mainly due to the stringent instructions for cephalosporin use that has been imposed by the Greek ministry of health since 1997. The observed increase in the use of macrolides may be explained by the availability of newer agents. Macrolides accounted for 32.6% of the total prescriptions, a percentage that was higher in comparison with other studies such as Molstad et al and Tzimis et al studies (21.9% of the total prescriptions and 22.5% of DDDs, respectively) [4, 14]. Macrolides were also used very frequently in the Anogeia Health Centre area, a neighboring county of Crete .
The previous prescribed overuse of second generation cephalosporines and macrolides is a well known problem worldwide that leads to a Streptococcus pneumoniae resistance resulting in serious not easily treated infections . Use of non recommended, more expensive, broader-spectrum antibiotics is frequent also in other studies . There is growing concern that using second generation cephalosporins and macrolides may lead common pathogens to develop antibiotic resistance to Streptococcus pneumoniae, Steptococcus pyogenes and Haemophilus influenzae [3, 25, 27]. Further antimicrobial resistance of S.pneumoniae to penicillin at a country level is often due to macrolides and beta-lactam antibiotics . A systematic review and meta-analysis on the effect of antibiotic prescribing in primary care on antimicrobial resistance showed that individuals that were prescribed an antibiotic in primary care for a respiratory or urinary infection developed bacterial resistance to that antibiotic . This increases the bacterial resistance to first line antibiotics, and has as a result an increased use of second line antibiotics in the community .
Although quinolones can be prescribed in Greece only under specific indications, their use was not rare in our study (2.2% of the total antibiotics), mainly for adult urinary tract infections, where quinolones held first position (42%). Interestingly not a single prescription of quinolones was documented in children. Comparing our findings with those previously reported for Greece , a fall in the percentage of sulfonamides and an increase of quinolones was observed.
In our study there was no registration of antiprotozoal medications and only one registration of a vermifuge medicine. This finding, compatible with a previously report for Crete , reflects the optimal hygiene level of the area. Regarding the seasonal distribution of the medical visits, the majority of the antibiotics were prescribed during the first quarter of the year. In Goosens et al study  the seasonal distribution in 32 European countries had a mean increase ≥30% in Southeastern countries in the first and last quarter of the year, while the mean increase in North countries was less than 25%.
The decrease of DDD, the lack of third generation cephalosporins prescriptions, and the lack of quinolones prescribed in children are indications that proper medical education and restrictions of prescriptions can help to improve antibiotic utilization in countries with high antibiotic use. On the other hand, the increase of prescribed broad-spectrum agents such as second-generation cephalosporins and quinolones stress that much more must be done.
Further this study showed that the primary indication for antibiotic prescribing was respiratory infections, both in adults and children. In Gjelstad et al study in Norway the respiratory infections accounted for the 60% of general practitioners' antibiotic prescriptions . Findings were similar in our study. In the Gjelstad et al study the antibiotic prescription has been shown to be influenced by various factors including type of infection, type of contact, being a general practitioner specialist, and years since medical exam . In a retrospective cohort study that took place in UK, adults had higher probability of receiving an antibacterial prescription for a respiratory infection visit compared to children . The antibacterial prescribing declined faster for younger patients than for adults after the implication of successful UK strategies for eliminating the antimicrobial use of medicines. Over the last several years a decline in antibiotics prescription in children has been reported in England  and this has been attributed to the significant reduction in GPs prescriptions. In our study a high antibiotic prescription for respiratory infections in children has been found, showing that further effort is needed to change GPs approach to childhood respiratory infections.
Implications to the Greek health care system
Today, Greece spends almost 9.7% of the GDP (Gross Domestic Product) for health expenses and there is much of discussion about the present economical state, and the medication cost in Greece as well as in Europe. Our study findings should be of interest to health policy makers who are in an extremely difficult financial situation at this time, which should generate more attention to the general practitioner/family medicine specialty that now represents only the 2% of doctors.
Strengths and limitations
This study has some limitations which should be reported. The study is a prevalence study so it does not give information about its change overtime. It is a small sample study so broad conclusions cannot be drawn. Large multicenter studies are needed in order to eliminate local factors that can influence the results. The study was conducted in only two rural areas of Crete, and therefore could not be representative of every Greek rural community, so the findings of this study could not be generalized for the whole Greek population. The use of over the counter medicines can not be excluded. The fact that we used a retrospective analysis of antibiotics in patients without interview may have as result an underestimation of the total antibiotic use.
Further the attempted comparison between adult and pediatric antibiotic use has several limitations as co-morbidity factors were not examined (diabetes, asthma etc) as well as a possibility of misestimating the results as DDDs are not recommended for children. On the other hand this is a real life study and it is among the few published regarding antibiotic use in primary health care in rural areas in Greece.