Article Text

Avoidable emergency admissions?
  1. Borge Lillebo1,2,
  2. Bodil Dyrstad1,3,
  3. Anders Grimsmo4
  1. 1Out-of-Hours Primary Care Centre, Varnesregionen, Stjørdal, Norway
  2. 2Norwegian EHR Research Centre, Department of Neuroscience, Norwegian University of Science and Technology, Trondheim, Norway
  3. 3Section for infection control, St. Olav's Hospital, Trondheim University Hospital, Trondheim, Norway
  4. 4Department of Public Health and General Practice, Norwegian University of Science and Technology, Trondheim, Norway
  1. Correspondence to Borge Lillebo, Varnesregionen legevakt Kjøpmannsgata 10-12, N-7500 Stjørdal N-7500, Norway; borge.lillebo{at}


Background Use of specialist healthcare services is increasing.

Aim To evaluate whether alternative healthcare services could reduce the need for admissions to specialist care hospitals.

Design Prospective observational study of emergency referrals for admission to specialist care.

Setting A single out-of-hours primary care centre (OPCC) in Norway.

Method Out-of-hours physicians registered their referrals for hospital admission and stated whether the admission could have been avoided given the availability of six other healthcare services.

Results Of 1083 registered encounters at the OPCC, 152 (14%) were referred for specialist care hospital admission. According to the referring physician, 32 (21%) of these referrals could have been avoided. The most eligible alternatives to such referrals were next-day appointments at a specialist outpatient clinic (11 of 32 referrals), or admission to a community hospital (21 of 32 referrals), or a nursing home (nine of 32 referrals). Respiratory (eight of 32 referrals) and gastrointestinal problems (12 of 32 referrals) were the most common among avoidable admissions.

Conclusions The use of specialist care hospital admission can be reduced if appropriate alternatives are available.

  • emergency care systems, admission avoidance
  • emergency care systems, primary care

Statistics from


The use of specialist healthcare services is expected to continue to increase. The Norwegian Ministry of Health has described this situation as non-sustainable, and has launched a white paper and a national health plan called The Coordination Reform.1 This reform encourages development of primary care services to ‘reduce the need for specialist health care services’. One proposal is admitting selected emergency patients to community hospitals as an alternative to specialist healthcare hospitals. Economic incentives, such as municipal cofunding of specialist care hospital admissions, are implemented to underpin this development.2 However, it is uncertain how many and what kind of patients could be offered alternatives to specialist care hospital admissions.

At present, the specialist healthcare hospitals in Norway are state-owned and managed through four regional health authorities. Regional authorities also manage community mental health centres. Private hospitals are small and sparse and need not be discussed further in this paper. Primary care services comprising general practitioners (GP), out-of-hours service, home care services, nursing homes and community hospitals are the responsibility of the local authorities (table 1).

Table 1

Description of Norwegian healthcare services that are mentioned in this article

Unlike in many other countries where emergency departments dominate the initial care of emergency patients,3 local GPs have had and still have an important role in initial care of emergency patients in Norway. A reasonable explanation for this manner of organising emergency care might be the distributed settlement, low population density and extreme climate and terrain in Norway, causing prolonged transportation to hospitals. In recent years though, more neighbouring municipalities have started to cooperate in larger out-of-hours healthcare districts, partially to improve the quality of the service and also as a response to recruiting problems.4

This study was conducted at one out-of-hours primary care centre (OPCC) comprising four neighbouring municipalities. The objective of the study was to evaluate whether future healthcare services could reduce the number of emergency admissions to specialist care hospitals. If so, what is the proportion of admissions, and what are the kinds of patients and problems that could be avoided, and what are the kinds of healthcare services that are likely to be the most useful alternatives to hospitalisation at a specialist care hospital?


We did a prospective observational study of emergency referrals for hospital admission from a single OPCC. For each patient referred for hospital admission, the referring physician indicated whether or not other care alternatives could have prevented the referral. Six alternatives were stated: keeping the patient for observation at the OPCC, providing the patient with intensive home care, scheduling a next-day appointment at a specialist outpatient clinic (OC), admitting the patient to a nursing home, a community hospital or a community mental health centre (table 1). These alternatives were chosen because they were mentioned in the new health reform,1 or they could be seen as natural extensions of current services.

The OPCC was staffed with one physician and one or two nurses who provide emergency primary healthcare to 29 000 inhabitants from four neighbouring municipalities. Transportation to the OPCC from home spans from a few minutes to 2 or 3 h. An ambulance takes approximately 30 min to travel from the OPCC to the specialist care hospital. The OPCC physician is usually a non-specialist, and he/she acts by regulations as a gatekeeper between primary and specialist care services. During office hours, patients consult their regular GP. The rest of the day, and during weekends, the GPs take turns at working at the OPCC. This organisational model for providing OPCC has been referred to as ‘Rota groups’.3

Data was collected between January and June 2011. This was done prospectively on a handy registration form by the physician on call at the OPCC. The registration method and outline of the form was according to the Audit Project Odense (APO) method.5 The included variables were based on a corresponding registration used nationally to monitor the use of out-of-hours services in Norway.6 We adjusted some of the variables to fit the specific goals of this study, and to make it easier for the physician to fill out the form continuously without interfering with clinical work. Age, sex and work shift (8:00–16:00, 16:00–22:00, 22:00–8:00) were registered. Main diagnosis (ICPC-2) was registered for patients whom the physician admitted to a hospital. Additionally, the physician had to state whether or not admission could have been avoided given the availability of the six alternative healthcare services. The physician was also encouraged to make a written comment on his opinion—that is, why the admission could or could not be avoided. The registration form was piloted during two work shifts to ensure that it was understandable and easy to fill out. Study size was chosen based on comparable studies.7 ,8

We introduced the registration form to all physicians working at the OPCC during the study period though it was voluntary to fill it out. The project was presented for the Regional Ethical Research Committee, and was considered to be a clinical auditing not needing further approval. Data was summarised manually, and the main results are presented as descriptive statistics.


The OPCC physicians registered 1083 patient encounters during the study period—37% of all patient encounters at that OPCC during this period. The characteristics of the registered patient encounters compared with all patient encounters at the OPCC during the study period, and compared with national figures are presented in table 2.

Table 2

Distribution of characteristics of registered patient encounters compared with all patient encounters at the OPCC during the study period, and compared with estimated national figures based on registrations from ‘The Watchtowers’ in Norway 20076

In all, 152 (14%) of the patient encounters ended up with an admission to a specialist care hospital. According to the physicians, 32 of these admissions could have been avoided. Community hospital admission was the preferred alternative to specialist care hospital admission. No physicians considered observation at the OPCC or intensive home care as eligible alternatives (table 3). Patients more than 74 years old constituted 138 encounters, 54 admissions (39% of encounters) and 13 avoidable admissions (24% of admissions). Among the avoidable admissions, only one child was registered—a 6-year-old with abdominal pain for which a next-day appointment at a specialist OC was the stated alternative.

Table 3

Avoidable admissions and preferred alternatives according to referring physician

The ICPC codes, and free text comments from the registration, conveyed some information about the clinical features of the avoidable admissions. Respiratory, gastrointestinal and cardiovascular problems (e.g. pneumonia, shortness of breath, abdominal pain, stroke) were the most common among the unavoidable specialist care hospital admissions. Except for cardiovascular problems, problems originating from respiratory or gastrointestinal organ systems were the most common among the avoidable admissions as well (table 4). Three admitted patients—for whom the admissions were unavoidable—were registered without a diagnosis.

Table 4

Organ system involvement for those patients who were admitted to a hospital

The GPs wrote 15 comments related to the avoidable hospital admissions. Four of the comments stated ‘possibly/possible’—thus weakening the GP's own decision. Other comments specified either the diagnosis (eg, ‘Food poisoning’ added to diagnosis ‘Abdominal cramps’), or the treatment (eg, ‘Needed intravenous administration of antibiotics’). Although this data was scarce, it may look as if the GPs expected municipal institutions to be able to administer intravenous treatment (antibiotics or fluid) or laxatives (for constipation), and to be able to observe patients with unspecific, non-vital complaints.

Patient encounters during day-time shifts were more frequently registered (44% of total day-time encounters) than afternoon (39%) and night shift encounters (19%). Patient encounters on Mondays were more frequently registered (52% of total Monday encounters) than Sunday encounters (28%) with remaining weekdays in between.

Although registered encounters were unevenly distributed with respect to weekdays, the admission per encounter rate did not vary much (13%–16%) except from Mondays with only 9% admissions per encounter. Avoidable admissions were more common during afternoon shifts (25 of 91 admissions; 27%) compared with day (five of 39 admissions; 13%) and night shifts (two of 22 admissions; 9%), while admission rates at night shifts were higher than day and afternoon shifts (21% vs 13% vs 13%).

Eighteen of a total of 27 OPCC physicians participated with registration of patient encounters during the study period. Their contribution to the data material varied with respect to the total number of registered encounters (range 7–354, median 31), admissions (range 0–46, median 4), and avoidable admissions (range 0–15, median 0.5). As a group, these GPs admitted 14% of their OPCC patient encounters, of which 21% could have been avoided. Two physicians registered considerably more patient encounters (48% of all encounters) than the other 16 physicians during the study period. These physicians had a lower admission percentage (13% and 12%) and a higher avoidable percentage (33% and 25%) compared with the remaining physicians (15% admission/encounter, and 14% avoidables/admission). The pooled result of all but the two most active physicians gives 563 encounters, 86 admissions (15% of encounters) and 12 avoidable admissions (14% of admissions). The median percentage of the individual GP's admission rates, and the median percentage of the individual GP's avoidable admission rates were both 13%.

Comparison of manually registered encounters with the number of encounters automatically recorded in the electronic patient record system indicated that 36 (3.3%) encounters were missing from the manual registrations.


To our knowledge, this is the first study that has included community hospitals as well as five other alternatives to admission to specialist care hospitals in a prospective evaluation of avoidable emergency admissions. This study shows that according to the GP on call, approximately one out of five emergency hospital admissions could have been avoided given the availability of other healthcare services. The most eligible alternative to such admissions was admission to a community hospital, but admission to a nursing home or a next-day appointment at a specialist OC were also frequently registered alternatives.

In our opinion, the design of this study—that is, a prospective judgement by the referring GP on call—is superior to other studies of avoidable admissions in which hospital physicians and/or expert panels evaluate the appropriateness of admissions retrospectively.8 ,9 A limitation of our study, though, was that we did not follow-up on the admitted patients and, thus, were not able to conclude whether any of the 32 patients would have suffered—or benefited—from alternatives to hospitalisation.

We stated six alternatives to hospital admission. It is possible that including more or less alternatives would have given other results. The GPs were not given a specification of what kind of services the alternatives to hospitalisation would provide, such as what kind of supplementary examinations and/or therapies could be done at a community hospital or a nursing home. Hence, community hospitals—as a rather new and unfamiliar healthcare service—might have been over- or underestimated as a proper alternative.

The results from this study heavily depend on the registrations of two physicians. These physicians had a lower admission percentage and a higher avoidable percentage compared with the remaining physicians. This might reflect differences in experience with OPCC work, but unknown or random factors may just as well also contribute.

Registrations of patient encounters were skewed with respect to both weekday and work shift, but in spite of these discrepancies the demographics were comparable with both local and national figures. One important difference, though, was that the national figures were based on all OPCC contacts of which 34% was handled by a nurse, or as a telephone consultation by the physician, while our registrations and the local figures only included patient–physician encounters (office consultations, home visits and call-outs). Another possible bias in our material was incomplete registration of patient encounters. Comparison with data from the electronic patient recording system indicated that 36 patient encounters were missing from the manual registrations. If all these missing encounters were avoidable admissions, our results would have been very different (admission rate 17%, avoidable admission rate 36%). If they were all unavoidable admissions, our results would have been slightly different (admission rate 17%, avoidables rate 17%), and finally, if they were all patient encounters that were not referred for hospital admission, our results would barely have been affected (admission rate 14%, avoidables rate 21%). We find it more likely that missing encounters were those that were not admitted to hospital, since ‘avoidable admissions’ was the expressed purpose of this study.

For comparison, the OPCC in Trondheim reported that 16% of all encounters in 2011 resulted in transfer of the patient to a hospital emergency department.10 The OPCC in Trondheim serves a population of approximately 207 000, and is the nearest OPCC to the OPCC at which this study was conducted.

Despite these limitations, we believe that our results can be generalised to other out-of-hours primary care settings with comparable organisational models such as ours. The fact that other studies of avoidable admissions have found similar results does, in our opinion, support this assumption.

Elwyn and Stott did a retrospective study of avoidable referrals from a single-handed general practice.7 They found that 58 (35%) of 168 referrals could have been avoided. Their results should not be directly compared with our results due to the major differences in study design, context and population. Other studies of alternatives to specialist care hospital admissions have found 23% and 24% inappropriate or avoidable admissions.8 ,9 These numbers were based on the opinion of hospital physicians, or expert panels consisting of hospital physicians and GPs. The study of Eikeland et al8 also included the opinion of the referring physicians (81 admissions), but did not report how many of these could have been avoided. Instead, they focussed on inter-rater agreement between the referring and the hospital physician on the necessity of admission, and found that they often disagreed. The authors concluded that establishing alternatives to emergency hospital admissions probably are of limited value if the referring physician thinks that an admission is necessary.

Furthermore, Eikeland et al8 found that most patients who were eligible for alternatives to admission were elderly people, and even that the relative share of alternative services was larger among the elderly. Our study also demonstrates that the elderly dominate the patient population among the avoidable admissions, but they also dominate among the unavoidable admissions. This discrepancy between the studies should be further studied.

Our data does not support any conclusions as to whether patients eligible for community hospitals also could be taken care of at regular nursing homes or vice versa. However, for eight of the avoidable admissions, both community hospital and nursing home were marked as alternatives to specialist care admission.

None of the GPs chose observation at the OPCC as an alternative to hospital admissions. This was quite surprising compared with the study by Eikeland et al8 which showed that an observation unit was considered the most relevant alternative to emergency admissions. A possible explanation could be that some physicians at the OPCC already used a spare bed at the OPCC to observe patients for some hours, so this alternative was perhaps already applied before even considering hospital admission. Similarly, ‘Intensive Home Care’ was not considered to be an alternative to hospital admission. In a meta-analysis of patients provided with hospital care at home, Shepperd et al11 found similar outcomes to inpatient care, but with a reduction in costs.

The number of avoidable admissions varied between shifts. This could be caused by registration bias related to the skewness of work shift registrations. Another explanation could be that the medical problems of patients presenting at the OPCC varies with time of day. ‘Watchtower’ figures have demonstrated that although most patients are dealt with during day time and afternoon shifts, the relative component of urgent and acute patients is larger during night shifts.6

A concern regarding avoidable admissions is the risk of making a wrong decision on which care alternative is the more appropriate for the patient. However, much research indicates that although primary care is associated with poorer quality care for diseases, it is also associated with better health for whole people.12 When the largest healthcare system in the USA reorganised during the 1990s, reductions in hospital utilisation accompanied by improvements in primary care were not associated with reductions in survival rates.13 Nevertheless, more research is needed to identify what kinds of patients would suffice—or even benefit—from non-specialist healthcare service alternatives.

This study provides additional evidence that emergency admissions to specialist healthcare hospitals can be reduced. In light of these results, there is reason to believe that the imminent health reform stands a fair chance of achieving some of its goals, including reducing the use of specialist healthcare services.

There is still much to learn about how these patients should be managed locally, and how the collaboration with specialist care services should be organised. Further research should be carried out to evaluate the clinical outcomes of patients eligible for alternatives to emergency hospital admission.


To the nurses and physicians at Varnesregionen Out-of-Hours Primary Care Centre.


View Abstract


  • Contributors All three authors participated in planning, writing and acceptance of the final paper. B Lillebo managed the data collection.

  • Competing interests None.

  • Ethics approval Norwegian Regional Ethical Research Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.