Since the start of the COVID-19 pandemic, some long-term trends in hospital care have been reversed (see Box 1). Most notably, between 2019 and 2022, the total number of hospital admissions in England fell – the first sustained reduction in hospital admissions since the early 2000s. This is despite increases in the number of staff working in NHS hospitals over the same period and a substantial backlog of people waiting for routine hospital care – currently standing at 7.3 million people.  

Although hospitals are still dealing with many effects of the pandemic, national policymakers are focused on increasing activity and restoring NHS services – including a headline ambition to deliver 30% more elective activity in 2024/25 than before the pandemic, and to reach 76% of A&E patients being seen within 4 hours by March 2024 (from 72.4% in January 2023). The first target to eliminate elective waits longer than 18 months by April 2023 was missed.  

If policymakers want more patients to be seen in NHS hospitals, they first need to understand why hospital admissions have fallen since 2019. Yet analysis of these changes and what may be driving them is limited. In this analysis, we explore patterns in hospital admissions between 2019 and 2022, drivers of recent falls in admission numbers and implications for policy. We analyse data from Hospital Episode Statistics (HES), which include information on all NHS-funded hospital admissions in England. We compare 2019 and 2022, using data on patients aged 18 and older. Our analysis focuses on trends in the number of admissions, total bed days and mean length of stay, including variation by NHS region, patient age and area-level deprivation (see Box 2 for more detail on our analytical approach).

Hospital admissions decreased while length of stay increased between 2019 and 2022 

In England, there were 800,000 (12%) fewer emergency and elective hospital admissions, excluding day cases, in 2022 than 2019 (Table 1). This is a major reversal of trends from before the COVID-19 pandemic. The percentage reduction was greater among elective admissions at 21%, than emergency admissions (9%).  

Also reversing pre-COVID trends, average length of stay rose from 7.3 days in 2019 to 8.3 days in 2022 overall. This pattern differed between emergency and elective admissions, with only emergency admissions seeing an increase in average length of stay (Table 1, Figure 1). Mean length of stay among admissions with at least 1 night in hospital decreased slightly from 5.2 days in 2019 to 5.1 days in 2022 for elective admissions (3% decrease). Emergency admissions, alternatively, saw an increase from 7.9 to 9.1 days (a 15% increase) in average length of stay. Median length of stay remained stable for both emergency and elective admissions. 

Overall, the total number of bed days recorded across all hospital stays decreased from 36.2 million in 2019 to 35.3 million in 2022, a reduction of 2.5%. This aggregate figure masks trends in opposite directions for elective and emergency admissions: bed days used for elective admissions decreased by 21%, while they increased by 1% among emergency admissions, meaning that emergency care took up an increased share of total hospital bed days in 2022. Moreover, the 2.5% decrease may partly be the result of a shift to treatment elsewhere in hospitals, where they are not counted as admissions, eg in Same Day Emergency Care settings, and these no longer being recorded in HES.

Note: mean length of stay is calculated among hospital admissions lasting at least 1 day.

Numbers of day cases, maternity admissions and regular day/night attenders also decreased between 2019 and 2022, as well as corresponding bed days, indicating that decreased emergency and elective admissions were not offset by an increase in other hospital activity.

Drivers of changes in length of stay and emergency admissions

Among emergency admissions, which account for over 80% of bed days, striking changes in length of stay took place between 2019 and 2022. There were around 562,000 (11%) fewer emergency admissions lasting 14 days or less in 2022 than in 2019 (Figure 2, Figure 3). In contrast, there were an additional 41,000 (8%) admissions lasting more than 14 days took place. The percentage of zero-day admissions – where patients are admitted and discharged on the same day – remained at around one-third of emergency admissions (1.7–1.8 million) in both years. 

The increase in length of stay was therefore driven by fewer shorter admissions and, to a lesser extent, by more longer admissions. This is also evidenced by changes in the total number of bed days: in 2022, admissions with stays up to 14 days accounted for fewer bed days overall than in 2019. But there were almost as many additional bed days for admissions with stays longer than 14 days, suggesting that fewer bed days for shorter admissions compensated for more bed days used for longer admissions. Patients with emergency admissions lasting longer than 14 days required an extra 1.5 million bed days compared with 2019 – bed days that were therefore not available for other admissions, including those with more minor needs or undergoing planned hospital care. 1.2 million fewer bed days were used for admissions lasting up to 14 days in 2022 compared with 2019. 

Although the available data do not establish causality, it is likely that increasing length of stay in turn contributed to the decrease in the number of admissions. Emergency patients staying longer in hospital likely added more pressure in hospitals, which already had high levels of bed occupancy. Without enough beds and unable to increase capacity by adding more beds, hospitals were forced to change admissions thresholds – ie not admitting some patients for shorter admissions who would have been admitted in 2019. High bed occupancy has previously been shown to reduce the likelihood of hospital admission.

Changes in length of stay for emergency admissions differed across age groups, regions and deprivation

Age

The number of emergency admissions decreased in all age groups (Figure 4). In contrast, average length of stay increased across age groups – and most substantially for older people. For patients aged 85 years and older, it increased from 10.8 to 12.5 days on average, compared with an increase from 4.7 to 5.1 days among patients aged 16–44 years. 18–44-year-olds occupied 11% fewer total beds than in 2019, while older age groups occupied more total beds This may indicate factors particularly relevant to older patients are contributing to longer hospital stays, such as higher health care needs (eg from COVID-19 or delayed care during the pandemic), and challenges discharging patients with further need for support.

Region  

Similarly, the number of emergency admissions decreased while the mean length of stay increased in all regions of England (Figure 5). However, there was considerable variation and some regions experienced much greater increases in length of stay than others. The South West saw a 26% increase in length of stay, from 7.6 to 9.5 days on average, while the smallest increases were observed in the Midlands (from 7.9 to 8.5 days). The change in total hospital bed days over time also varied across regions: the greatest increase between 2019 and 2022 was in the South West (8%) and East of England (5%), while most other regions saw total bed days hold steady or fall. The extent of the increase in the South West is surprising, because this region already had higher bed occupancy levels than most other regions in 2019–20. Data quality issues could play a role in these differences, but regional variations may also reflect differences in local needs, previous service capacity and service configurations. 

Deprivation level 

Patients living in more deprived areas experienced a greater fall in admissions than those in wealthier areas (Figure 6). However, changes in mean length of stay were similar across deprivation levels, with an increase of around 14% across all deciles (although mean length of stay remained lower in 2022 among the most deprived deciles). Overall, total bed days decreased by up to 3% among patients living in the three most deprived deciles, but increased by up to 3% among less deprived deciles.  

These patterns by deprivation are not explained by people of different ages living in areas with different deprivation levels: we see the same trend of more pronounced decreases in emergency admissions and bed days within each age group. In addition, greater reductions in emergency care among people living in more deprived areas were not offset by smaller reductions in elective care. Although the patterns by deprivation are less clear than for emergency admissions, the largest reductions in number of elective admissions and elective bed days occurred in the most deprived decile.

NHS trusts 

Variation in length of stay between trusts widened over time, with mean length of stay ranging from 6–50 days in 2019 compared with 6–48 days in 2022. Figure 7 shows 85 of 105 NHS trusts (81%) reported an increase in mean length of stay (by >0.5 days between the 2 years), while it decreased (<-0.5 days) or remained stable in 19% of trusts.

Note: hospital trusts are pseudonymised in HES, so we were unable to include provider names.

Our analysis of national data does not definitively establish the drivers of increasing length of stay and the decreasing number of shorter admissions, which likely vary across hospitals and different areas of England. But it does point to several potential factors that may be contributing to the patterns observed.  

• Fewer patients presenting to hospital. One possible explanation for fewer shorter stay admissions could be reduced need for care among the population, or that people are less likely to seek emergency care for minor conditions than pre-COVID-19. However, the data suggest this is unlikely. The number of attendances at major A&E departments has not changed materially, at around 16.2 million in 2019 and 2022. The percentage of these attendances leading to admissions decreased slightly from 30% to 27%, suggesting increased admission thresholds is a more likely explanation than shifts in patient needs or behaviour. 

• Patients with minor conditions being cared for outside hospital wards. New care models trialled to reduce some hospital admissions may be contributing to decreases in shorter admissions, by providing care to people with less acute needs in alternative settings. These models include ‘virtual wards’, using remote technology to care for people at home; urgent community response services, enabling rapid multidisciplinary support at home for patients with complex needs, and Same Day Emergency Care–. However, different care models have been implemented around the country in a patchwork manner, meaning that patients not admitted in some places or with some conditions may not have been treated in hospital at all. Evidence on whether approaches like these can deliver durable reductions in hospital care use is also mixed.  

• Barriers to managing patient flow:

  • High bed occupancy. Patients may be staying in hospital longer due to inefficiencies in patient flow, leading first to delays in assessing, diagnosing or treating patients, and second, to delays in discharging patients once they are medically fit to leave hospital (eg due to limited capacity in social and community care). High bed occupancy is a symptom of pressures on services, leading to delays along the care pathway that may contribute to longer stays in hospital.

  • COVID-19. The continued presence of COVID-positive patients has meant hospital services working with infection prevention and control protocols, which can reduce bed capacity. The number of overnight hospital beds (around 100,000) and bed occupancy levels (around 92% in October to December) was similar in 2019 and 2022, but some unoccupied beds were reserved for COVID-19 admissions (around 0.3% of all adult general and acute beds in April 2023). This affects other aspects of patient care including patient flow.

  • Delayed discharges. The number of patients experiencing discharge delays almost doubled from 8,000 in October 2020 to 14,000 in February 2022 – the highest number on record. This was due to limited capacity in social care and NHS community services to accept discharged patients. National data do not make it clear whether increases in length of stay can be attributed to changes before or after patients are ready for discharge.  

• Increase in patients who are more unwell. Patients admitted in emergencies in 2022 may be more ill, on average, than before the pandemic and therefore require longer periods in hospital. This could be due to ongoing additional burden of COVID-19 infections, increased waiting times to access elective care and potentially secondary effects of COVID-19 infections and pandemic restrictions. Around 4% of hospital beds in December 2022 were occupied by COVID-19 patients, who are more likely to have longer lengths of stay than other patients.  

Overall, it is likely that the trends we describe are driven by a combination of some factors, as well as other factors not explored here (eg staff sickness and low morale following the pandemic). 

Compared with 2019, in 2022 the number of patients being admitted to hospital fell by 12%. Elective admissions were 21% lower and emergency admissions 9% lower. A likely key driver of the reduction in emergency admissions is an increase in length of stay: as hospital stays lengthened and bed capacity was stretched even further, hospitals had little choice than to reduce admissions, with knock-on consequences for A&E services and the elective backlog. Our results highlight several implications for policymakers and service leaders. 

Meeting the NHS’s ambitious targets for elective and emergency care recovery will be challenging without successfully increasing hospital activity. Although the most recent figures indicate that elective inpatient admissions are almost back to March 2019 levels, the NHS needs to achieve the 30% increase in elective activity by 2024/25 outlined in national plans for elective recovery. A first national target to eliminate elective waits longer than 18 months by April 2023 was missed, underlining the scale of the challenge. Emergency admissions are still some way below 2019 levels. However, this does not necessarily imply NHS staff are working any less intensively: staff are dealing with very similar numbers of patients in beds and pressures in hospitals may mean they are unable to work as efficiently (eg more effort may be required finding beds and managing discharge). 

Overall, our analysis suggests that hospitals have increased admission thresholds in response to increased pressures on beds. This is a consequence of a system which, in 2019, was already under significant strain, with little or no capacity to deal with surges in demand. It is unclear what is happening to the roughly 560,000 patients not admitted for shorter emergency stays, who would have been admitted in 2019. Some may be accessing care in community services or enhanced A&E units rather than being admitted. But the extent to which conditions are being left untreated is not clear. Appropriate care may be being rationed in unexpected ways. This has implications for people’s health and other services (eg increasing GP services’ workload). It also presents some questions about longer term impacts on patient experience and outcomes, as well as future health care needs for conditions with present unmet need.  

The reversal of trends in admission numbers and length of stay since the pandemic strengthens the argument that the reduction in hospital beds over the past few decades went too far, and that additional capacity is needed. Plans to increase hospital bed stock by 5,000 in 2023/24 are a necessary near-term response, but further increases are likely needed over the long run to account for growing health care needs: we previously estimated that 23,000–39,000 additional beds may be needed by 2030/31 just to maintain existing service levels. Making this happen would require substantial additional investment and expansion of NHS staff. Strengthening services outside hospitals – most notably social care – will also be needed to ease congestion within hospitals, alongside other improvements to hospital flow to help make the most of existing resources.  

Our analysis found that changes in admissions and length of stay varied across the country. The drivers also likely differ from place to place. More analysis is needed – at national and local levels – to understand what is happening and the actions needed. Local leaders should explore the factors driving up length of stay in their hospital trusts, particularly among groups with the highest increases such as older patients. While some of the increase may be appropriate (eg due to more complex health needs), there may be opportunities to better coordinate inpatient and discharge care, and improve flow for the benefit of patients and system capacity. Understanding regional variations – including, for example, how the South West and East of England increased emergency bed days despite high rates of bed occupancy pre-COVID-19 – may generate useful learning for accelerating NHS service recovery. 

Lastly, there was a greater decrease in emergency admissions for patients living in more deprived areas, leading to a reduction in bed days for emergency admissions, while emergency bed days increased in less deprived areas. Moreover, decreases in bed days for elective admissions were greatest in the most deprived decile, indicating large reductions in hospital care used by people living in the most deprived areas compared with the wealthier areas. This raises important concerns about worsening equity in access to hospital care since the COVID-19 pandemic. These trends run alongside higher pressure on GP services and larger funding cuts to local authorities in more deprived areas – and go against national leaders’ aspirations to reduce health inequalities and tackle the NHS’s care backlogs in ways that focus on meeting the needs of the most deprived 20% of the population. In the wake of COVID-19’s unequal impact across communities, more priority needs to be given to tackling inequalities in access to care. 

Much changed between 2019 and 2022. A global pandemic disrupted people’s lives, causing huge damage to population health and demanding rapid and extensive changes in health and care services. This analysis suggests hospital care also changed.  

Patients are staying longer in hospital and this is leading to reductions in shorter emergency hospital admissions and in elective admission numbers. These developments represent departures from long-term trends in NHS care over previous decades.  

NHS goals for the recovery of emergency and elective care are predicated on increasing hospital activity – if hospitals are not able to do this, these ambitious targets may not be achievable. It should therefore be a priority for policymakers to gain a fuller understanding of what is contributing to these changes in hospital care – including drivers of increasing length of stay, and the knock-on effects on admission thresholds and elective admissions.