The number of visits to EDs decreased during the follow-up of the interventions. This decrease was most prominent in the youngest age groups. Especially, the proportions of recorded infectious diseases (Groups A, B and J) decreased. Particularly, diagnoses related to mild infections of respiratory airways decreased. Interestingly, the effects of interventions on the prevalence of different injuries varied, although generally the proportions and absolute numbers of injuries increased. The proportions and numbers of symptomatic diagnoses increased.
The implementation of the ABCDE-triage system for assessing patient acuity at Peijas combined ED reduced the number of patient visits to GPs of the ED by 8 % , closing the minor suburban ED brought the reduction in visits to 15–20%  and further applying “reverse triage” brought the total reduction in visits to about 25% . No increased mortality was found . Providing enough information about the ED changes to the population proved to be important for the success of the implementation of the present interventions [2,3,4].
The decreased rate in the use of primary care EDs in the youngest people (0–19 years) is understandable. There are earlier reports suggesting that primary health care ED services are often used by the younger age groups [11, 12]. Especially low acuity visits to EDs seem to be a feature of very young age groups (< 10 y) and late teenagers (18–19 years) . Interestingly, social deprivation does not seem to influence this pattern of ED use . The reasons for this are many. According to a survey study with ED patients, young age groups may differ in their expectations regarding the purpose of out-of-hours services and accessibility and they may have other objectives than plain clinical urgency when they seek help . Furthermore, in a multicenter survey of patients from an urban health region, distance to a specific ED was the most important reason for choosing that ED suggesting that convenience factors play a significant role when deciding to use ED services . There is support for this view from other studies analyzing primary care out-of-hours calls and visits concerning child patients  and young adults  as well as from qualitative studies regarding treatment of small children in out-of-hours services . Nevertheless, the use of ED services in the youngest age group is strongly regulated by their adult parents. The adults did not seem to reduce the number of their own visits as strongly. Based on this and former data [14,15,16,17,18] the adults may change their estimation of the acuity of the health problems of children and youngsters more readily than of their own, if the inconvenience of visiting the ED is increased, with longer waiting times and prolonged distances to reach the ED. Altogether, previous findings have shown that if restrictions of access to primary health care EDs occur, the youngsters, who use these services a lot, reduce their ED visits more than other people.
Where did the patients go if they did not come to the primary care ED? Some patients were redirected to a single speciality ED after applying ABCDE-triage  and there may have been some marginal spillover to private sector primary care after closing the minor suburban primary care ED . Some patients were redirected to both a single speciality ED and private sector primary care after starting the “reverse triage” . However, the number of patients who were redirected to other practices was less than the reduction in the number of visits to the primary care ED [2,3,4]. Neither was there a spillover to public office-hours primary care [5, 18]. As there was no change in the secondary ED functions during the follow-up period the people of Vantaa might have changed the way they estimated the acuity of their health problems.
Access to individual patient data would have given detailed information about the safety of the combination of ED-interventions that were applied. We had only mortality data, which gave a crude but definite estimation of the safety of the studied interventions [3, 4]. The overall mortality in Vantaa increased during the present follow-up period because of ageing of the population . However, in the studied age groups 20–64 and 65+ the mortality tended to decrease . Only a slight plateauing of the decreasing monthly mortality was observed in the oldest age group (65+) after 2008, the year of the implementation of “reverse triage”,  but analogous plateauing was observed in the mortality of the whole Finnish 65+ age group in the period 2008–2014 . Thus, no lethal safety hazards were found in this analysis.
The finding that the proportions and numbers of simple infections in the ED decreased is in line with the aims of the interventions applied [2,3,4]. We knew that at least about 30% of the diagnoses done in the present kind of primary care ED system and office-hours primary care were the same and that office-hours primary care might therefore have provided better continuation of treatment for these patients than the ED . Furthermore, when diagnoses in EDs and primary care doctor driven emergency systems have been compared, a higher prevalence of mild infections in primary care doctor driven emergency systems and a higher prevalence of injuries in EDs have been reported . This injury-focused activity in EDs has also been described elsewhere in all age groups [23, 24]. Thus, the present triplet of interventions seemed to shift the functions of the studied primary care ED towards the form of a standard specialized health care driven ED. Whether these low-acuity primary care services should then also be provided to the population out of office-hours is another question .
This was a retrospective study considering primary care EDs. As this study was purely register-based the subjects were not aware of their participation in the study. The present result reflects real clinical activity in this respect. As a confounding factor, electronic reminders were introduced in the electronic patient information system in 2008 to enhance recording of diagnoses and that may have altered the observed proportions of different diagnoses during the present study . For example, this intervention explains at least partially the observed increase in symptomatic diagnoses (IDC-10 group R) during this study .
As a limitation, we should have been able to compare our results to a control city with a similar office-hours primary health care, demography, and size. This would have strengthened our conclusions. However, such data from another city were partially available for comparison only for the first two interventions [2, 3]. Data about possible changes in patient material or changes in ways to manage practices and diseases were not available. These factors have a considerable effect on changes in the number of visits to GPs. Data concerning these putative changes could have been obtained if we had had access to the patient information of individual patients. This would have allowed us to follow individual patients and their fates instead of plain numbers of visits. Furthermore, we would have needed data about how visits to the primary care nurses developed during the follow-up as the present type of interventions may shift patients from physicians to nurses . Unfortunately, we were unable to obtain access to these data.
To our best knowledge, the health policy or system in Vantaa was not changed during the follow up but, naturally, we were not able to control all putative secular trends that can account for the observed reductions. For example, the population of Vantaa aged rapidly during the follow-up period . This was the most obvious of these social trends but ageing of the population should not decrease the use of EDs. Cost savings were not observed as the health care costs increased linearly throughout the follow-up period .