Consultation is a common and important aspect in the departmental function of emergency medicine.1 2 An emergency consultation occurs when an emergency physician contacts another specialist (consultant) within the hospital to continue care for patients when their medical problems require a service outside the scope of practice of the emergency physician. In many North American hospitals there are different types of consultation which an emergency physician can request: (1) consultation for admission (most common); (2) consultation for opinion only (where the patient can be discharged but an opinion is needed for specific investigations or outpatient arrangements); (3) consultation for treatment or special procedure (where a consultant assists with management of a specific problem); (4) consultation for “transfer of care” (where a consultant takes over care of the patient); (5) consultation for outpatient referrals (patient sent for outpatient follow-up).

An emergency consultation differs from an outpatient referral in that an in-hospital consultation is an assessment followed usually by same-visit recommendations or interventions by the consultant who has a specific level of expertise.3 The standard definition of consultation describes emergency consultations as requests for assistance while the patient remains under the care of the emergency physician.4 However, an outpatient referral is less urgent and may request for a physician to assume direct responsibility for the management of the patient for a specific period of time or to provide expert opinion to assist the patient’s family physician in ongoing management.3

A study by Cortazzo et al provided evidence that the overall rate of consultation and referral was 39.9% in a military emergency department (ED) in the USA.4 Reports from the medical literature suggest that the proportion of physician consultations in the ED varies between 20% and 60%.5 In many countries ED overcrowding is a challenging but important health care issue,6 and there is evidence that lack of timely consultation is a leading cause of ED overcrowding, at least in Canada.7 Delays in consultation response times may be due to specific barriers including: physician inconvenience, competing priorities, lack of financial incentives, shortage of physicians and failure to enforce rules regarding policies on consultant coverage.5 Other studies report that these times vary depending on the individual consultant, the specialty consulted and the acuity of the patient’s condition.8 Guidelines from professional organisations suggest that a reasonable response time of consultants to the ED is between 30 and 45 min, although it is primarily dependent on the current needs of the patient.9 A survey of Canadian ED physicians showed that some physicians will overstate certain aspects of their patient’s clinical picture in order to obtain a quicker consultation with less conflict from other physicians.10 The input-throughput-output model conceptualises ED overcrowding and recently the model has extended to include inpatient hospitalisations;11 consultations would influence both throughput and output relationships.

Poor communication between physicians and delayed consultation response times ultimately affect the safety and quality of patient care. Problems with the consultation process in the ED have the potential to create life-threatening situations for patients and legal risks to health care providers.5 9 Guidelines suggest that conflicts should be resolved between physicians and consultants according to hospital policy and should never jeopardise patient care.9 Past studies also show that educating both the consulting physicians and consultants, or implementing a strict policy involving authority figures (eg, academic or clinical leaders) can effectively reduce response times and improve communication between physicians.12 13

While previous research has uncovered the importance of consultations in the ED and difficulties involved in this process, no systematic review related to this topic has been published. This review summarises the literature on the use of consultations, as well as interventions applied to influence consultation rates in the ED.

The purpose of the systematic review was to provide a descriptive overview of all studies examining the proportion of consultations either within all patients seen in the ED or a specific subpopulation of ED presentations. Furthermore, the review gathered evidence on interventions applied in hospitals to reduce or control ED consultation rates to a specialist.

RESEARCH QUESTIONS

Research questions answered by this study were:

1. In ED patients (population), what proportions of all patients receive consultations (outcome) to a specialist following ED assessment (test)?

2. In special ED groups (population), what proportions of patients receive consultations (outcome) following ED assessment (test)?

3. In ED patients (population), what proportions of those patients consulted are eventually hospitalised (outcome)?

4. For patients seen by ED physicians (population), can education, programmes or policies (interventions) reduce the proportion of patients receiving a consultation (outcome) compared with regular practice (control)?

Criteria for considering studies for this review

To be included in this review, studies must have focused on ED cases and specifically on patients receiving consultations. In many parts of the world the term “consultant” is often used to denote the specialist in a particular field (including emergency medicine). In this study we used the term “consultant” to refer to consultants who were not emergency physicians.

Studies reporting on the proportion of consultations in the ED or examining interventions to influence ED consultations were considered for inclusion. Eligible study designs included randomised controlled trials, quasi-randomised trials, before-and-after studies, cohort studies, prospective and retrospective studies. No restrictions on publication status (abstract vs manuscript) or year of publication were applied.

Search strategy for identification of studies

The following electronic databases were searched using standardised search terms: MEDLINE, EMBASE, PUBMED, SCOPUS, SCIRUS, Cochrane Library, Web of Science and Health Star (1966–May 2007). Various search terms were used such as “consult”, “referral”, “second opinion” and any other similar synonyms. In addition, grey literature searches included SAEM abstracts (2001–2007), CAEP abstracts (2001–2007) and Google Scholar. Finally, reference lists of five relevant studies were checked and experts in the area of interest were contacted. As one final effort to identify papers, we performed a cited reference search on Web of Science for all included studies. The search is considered updated to May 2007.

Selection of studies for inclusion

A two-step selection process was used for study inclusion. The initial search results were first screened by one reviewer (RL). Relevant articles were then retrieved in full. Two reviewers (RW, BHR) independently reviewed the full reports for final inclusion. Five studies in which disagreements were observed were reassessed by the reviewers to resolve disagreements.

Inclusion criteria

In order to be included in this review, studies must meet the following criteria:

• Study population:

• 1. Did the study include patients presented to the ED?

  2. Did the study deal with patients with acute problems?

• Study intervention: Did patients get consulted?

• Outcome measures: Did the study consider one or more of the following outcomes?

• 1. proportion consulted;

  2. proportion admitted/hospitalised;

  3. patient satisfaction;

  4. time in ED;

  5. occurrence of any adverse events;

  6. long-term outcomes.

Exclusion criteria

Studies were excluded if they reported on patients presenting outside the ED, who did not meet one or more of the study exposures, or did not present one or more of the specific outcomes. Excluded studies consisted of editorials, physician surveys, opinion-based studies commenting on the art of consultation and studies describing consultation patterns (see Appendix A).

Quality assessment

Several types of measurements were used to assess eligible studies such as the Jadad scale (randomised controlled trials, controlled clinical trials)14 and the Newcastle-Ottawa scale (cohort studies and before-and-after quality assessment). Missing information was sought through communications with authors.

Statistical analyses

Descriptive statistics are provided. For intervention studies, effect sizes were calculated. Weighted differences (WD) were calculated for each continuous variable and reported using 95% confidence intervals (CI). Odds ratios (OR) were calculated for each dichotomous variable and reported using 95% CI. No pooled results are reported.

RESULTS

Study selection

Using the search strategy described, 15 549 citations were identified of which 54 were found to be potentially relevant (fig 1). The full reports were reviewed independently by two reviewers and nine were initially included in the review (κ = 0.712). Disagreements (n = 5) were resolved by repeated evaluation; none required third party adjudication. After reassessment, three recently published abstracts were included in the review.13 15 16 Final agreement was κ = 1.0.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Search results.

Description of studies

Of the 12 included studies, 7 reports were published in journals, 2 were published in academic reports and 3 were abstracts. Four studies reported the proportion of patients receiving a consultation for all ED presentations, 6 reported the proportion of patients receiving a consultation for special populations of ED presentations and 2 examined interventions to influence consultations practices. Detailed information on each study is provided in tables 1, 2 and 3.

Years of publication

The majority of studies describing ED consultations were identified from older literature (1982–2003) except for two,15 16 while interventions applied to reduce or control ED consultations were more recent (2000–2004).

Country of publication

Of the 12 included studies, all but one was published in North America.

Designs

Descriptive ED consultation studies used a mixture of retrospective and prospective designs. The two intervention studies described the before-and-after intervention method. No randomised controlled trial studies were identified.

Setting

Most studies were performed in large urban hospitals. Five studies were conducted in teaching hospitals and two in military hospitals.4 15 17^–20 One study was conducted in a suburban community hospital.8

Consultation proportions

The number of patients reported in the studies varied from 91 to 4627. The proportion of patients consulted for all ED presentations varied from 20% to 40%,4 8 15 16 with the lowest at the suburban community hospital and the highest at urban teaching hospitals. The differences in samples size of the studies should be taken into consideration when looking at the variances in the proportions of consultation. The proportion of patients consulted at a paediatric ED was 32%.21 For special populations of ED presentations, two studies reported the proportion of patients receiving psychiatric consultations to be 5%.22 23 Within a sample of psychiatric patients presented in the ED, 50% of patients received a consultation.18

Hospitalisation proportions

The proportion of hospitalisations for those patients receiving specialist consultations was highly variable depending on the patient subpopulation studied and the type of hospital. Cortazzo et al4 reported that 87% of all consulted patients were admitted, and Woods et al15 reported a proportion of 54%. Curry and Wang16 reported an admission rate of 68% for the highest acuity site and 64% for the lowest acuity site. Among patients receiving psychiatric consultation, one study found that 57% of consulted patients were eventually admitted,22 while another reported a 31% admission rate. In the study population receiving geriatric consultation, 64% were admitted to hospital.20

DISCUSSION

This systematic review attempted to synthesise the best available evidence regarding consultations in the ED. Using a comprehensive and unbiased search of the published and unpublished literature, this study identified only 12 observational studies involving ED consultations. These studies involved more than 13 000 patients and the results of this review provide valuable insight regarding the role of consultations in emergency medicine. First, based on the available data, it appears that consultation research has virtually ceased since the mid-1990s. This is surprising, since evidence suggests that consultations are an important aspect of the delivery of care in the ED and may contribute to disposition delays and to the current overcrowding crisis in EDs.23 24 Second, all but one study were conducted in Canada or the USA. These findings therefore reflect consultation practices in North America, and consultation differences probably exist in other health systems. Third, this study shows that the proportion of patients who receive a consultation varies based on the setting and the types of patients from 20–40% overall to 5% for psychiatric consultations. Furthermore, there is some evidence that consultation rates at urban tertiary care centres were higher than in other locations (28–40% vs 20%).4 8 15 16 All studies reporting hospitalisation for patients receiving consultations in the ED indicate a high percentage of admissions among consulted patients.4 16 20 22

Moreover, there is heterogeneity with respect to the proportions of consultation based on service (5% for psychiatric patients and 32% for paediatric populations21^–23); multiple factors are probably responsible. There are also few studies examining interventions to improve consultations in the ED (n = 2) with diametrically different results. One study reported an increased length of stay (median duration 1 h) and more return ED visits from patients (9% to 21%) after a change in consultation policies from “routine” to “selective”.19 Another study showed a decrease in median consultation times (1.57 h) after a policy mandated by a university administration.13

Limitations

This review has several limitations. First, there were many potentially relevant studies that were excluded in the review as they were opinion-based, not primary research studies, or descriptive (fig 1 and Appendix A). Second, heterogeneity of setting, design and outcome measures precluded meta-analysis of consultation practice or comparisons of consultation proportions. Third, few comparative studies of interventions on ED consultations were found indicating possible publication bias. Furthermore, almost all included studies were conducted around 10–15 years ago in Canada or the USA thus limiting the generalisability of our findings. Finally, the literature in this area fails to report comprehensive outcomes and examine the role of consultation on ED overcrowding which further limits the value of the current scientific evidence.

Implications

This systematic review shows that consultation practices in the ED are seldom studied so more research in this field is urgently needed.15 This includes three specific types of research. First, a series of studies is required in a variety of ED settings, especially non-North American locations, to examine the variations in consultation proportions. Second, further research on the appropriateness15 and timing25 of consultation is also required. Finally, studies are required to examine the role of changes in consultation practices on ED overcrowding, quality of care and patient outcomes, and also the impact of ED overcrowding on consultation practices by emergency physicians and other specialists. Intervention studies should focus on increasing the efficiency and effectiveness of consultations by consult deferral, improving consultant timeliness and chronic disease management (eg, congestive heart failure, chronic obstructive pulmonary disease, asthma). Other interventions include educating ED staff and consultants and implementing a consult programme or policy with strict enforcement. Better designs for studies should be used such as a prospective analysis for ED rates and a comparative analysis for intervention studies. Intervention studies describing setting characteristics, study design and standardised outcomes will increase the understanding of the relative effectiveness on influencing ED consultations.

In conclusion, consultations in the ED are frequent and, in some studies, are requested in 20–40% of patients.4 8 15 16 Consultation appears to depend on the setting (teaching or community hospital; urban or rural area) and the ED census; however, this area has been relatively infrequently studied. There are few published studies on interventions to reduce or control ED consultations (n = 2). This review found that most published literature is descriptive in nature and of methodologically low quality (prospective and retrospective analysis). In addition, many studies report on consultation proportions for a specific subpopulation in the ED (eg, psychiatric, paediatric), thus limiting their relevance. Further research is therefore needed to evaluate ED consultations and institutional level changes to impact their effect on overcrowding.