Article Text

Download PDFPDF

Are rising admission thresholds good medicine?
  1. Adrian A Boyle1,
  2. Ellen J Weber2
  1. 1 Emergency Department, Addenbrookes Hospital, Cambridge, Cambridgeshire, UK
  2. 2 Department of Emergency Medicine, University of California San Francisco, San Francisco, California, USA
  1. Correspondence to Dr Adrian A Boyle, Emergency Department, Addenbrooke’s Hospital Cambridge University Hospitals NHS Foundation Trust, Cambridge, Cambridgeshire CB2 2QQ, UK; adrian.boyle{at}

Statistics from

Like many emergency physicians of a certain age, we find ourselves nostalgic for a time when it felt easier to admit patients into the hospital than now. Many of us now find ourselves feeling that we discharge many patients from the ED who we would have admitted years ago. It turns out that this hunch is correct. In an enormous retrospective database study, Steven Wyatt et al 1 examined admission thresholds across 47 EDs in England over a 5-year period. They used a simple regression model to adjust for known predictors of admission, including robustly recorded items such as age and sex, and less well recorded items such as diagnosis.

Initially, the conversion rate (the number of admissions divided by the number of attendances) looks similar across the years, but when adjusted for various measures of acuity, there is a 3% reduction in admissions over the 5-year period. This reduction occurs across all patient groups, but is greatest for adult walk-in attendances and least for children arriving by ambulance.

While the effect sizes are small, the very large study size allows the authors to estimate that 137 000 hospital admissions were avoided in the selected hospitals in 2015, if the 2010 admission thresholds had been applied. The authors estimate that this has saved around £65 million in 2015, although it is not clear whether there are extra costs in other parts of the system.

The study uses multiple hospitals, representing about a third of all NHS hospitals in England, so the generalisability is excellent. It is a plausible result, confirming what many of us have suspected. However, as with all database studies, some of the change may have occurred from better coding and automation of electronic patient records. The authors consider this, but point out that there was no meaningful change in fiscal incentives, which usually drive coding improvements during the study period. There was also an even gradient each year in the reduction of admissions, which argues against a recording bias. There has also been a substantial increase in the number of consultant emergency physicians in the UK, and it is plausible that these senior doctors are confidently and safely discharging many more patients home. Certainly, ambulatory care units, where patients can avoid an admission by prompt access to an inpatient specialist in an extended outpatient clinic setting, provide a safe alternative to admission for many patients. While it might be tempting (for some observers) to attribute a lower admission rate to what is thought to be a rising number of low-acuity visits, this was accounted for by acuity adjustments. Further, based on our daily experience, it certainly seems as if our patients overall are older, sicker and more complex than previously, and if anything we would expect more admissions.

Perhaps the best explanation is that, over the years, and world over, EDs have experienced increasing pressure not to admit patients to hospital. Some of these reasons are clearly for the benefit of patients—separating patients from family, exposing them to hospital-acquired infections and other iatrogenic complications, engendering risk for delirium in older patients. However, the pressures are also economic: hospital admissions are the most expensive part of any ED visit, so avoiding admissions saves costs to the system. Where exit block is a problem (and where it is not?), each admission contributes to the exit block of the next patient, also resulting in second thoughts about whether the patient can be discharged.

In response to these pressures, emergency physicians have upped their game considerably from the early days. No longer it is sufficient to stabilise patients and then admit them for the definitive diagnosis. This made sense in the ‘old days’ when hospitals were not full, and we thought any serious illness required admission. But we have evidence now that some of these patients can be managed as outpatients. As a result, we have had to get better at diagnosis, risk stratification and prognostication. For patients ultimately sent home with such serious maladies as Deep Vein Thrombosisor Pulmonary Embolism requiring anticoagulation, pneumonia requiring antibiotics and close follow-up, or syncope requiring home ECG monitoring, we have entered the disciplines of patient education, case management and social work. Additionally, we have become short-term inpatient physicians, obtaining D-Dimer, CT and troponins, performing serial abdominal and neurological exams, and monitoring response to intravenous antibiotics in our Clinical Decision Units. For patients at the end of life, we have begun discussions and treatment for palliative care, lowering resource use earlier in the patient’s admission. All of these have required an ever-expanding knowledge base, more interaction with subspecialties and an increased tolerance for risk.

However, while rising admission thresholds probably represent a saving to the health economy (and perhaps demonstrate the versatility of our specialty), it is not clear what the effect is on important patient outcomes, such as ED and hospital length of stay, intensive care unit admission, patient reattendance and mortality. A recent US study by Obermeyer and colleagues, for example, found, even after accounting for acuity, EDs that admitted fewer patients had a higher rate of mortality for patients discharged from the ED, than departments with a higher admission rate. Patients visiting hospitals with higher charges were also significantly less likely to die.2

There is no doubt that many patients do not benefit from hospital admission. And some indeed are harmed. However, the answer surely is not to avoid hospitalisation for those who need it, but rather to improve inpatient care. As we have seen with many espoused doctrines in medicine (take pain medication, for instance), we must be alert for signs that the pendulum is swinging too far in a direction that harms patients, and restore the balance.



  • Twitter @dradrianboyle

  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally 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.

Linked Articles