Skip to content
Long read

Lessons from the 2000s: the ambition to reduce waits must be matched with patience and realism


The NHS had a problem with rising waiting times for elective (planned) care before the arrival of Covid-19. In the main though, the declines in performance were fairly slow with the exception of repeated winter crises in A&E. Promises of additional staff and money from the government during the 2019 general election also held the prospect of stopping and then reversing the decline.

Covid-19 has changed all that. The number of people waiting for treatment has grown dramatically as has the length of time they wait. Measures to prevent Covid-19 infection will continue to reduce productivity and hamper efforts to ramp up activity. Staff exhausted by Covid-19 will also need time to recover before confronting the new waiting-time challenge. The backlog will be here for some time.

In the relatively recent past, the NHS made dramatic inroads into waiting times. In the 2000s, the NHS first  reduced waiting times in A&E (the four-hour standard), and then did the same for planned and routine treatment (the 18-week standard) and also delivered a range of tighter targets for other services like cancer. After reaching these new standards, the NHS then maintained them until the slow drip, drip of rising demand met the spending (and staff) restrictions of the austerity years.

It might be argued that this earlier transformation did not have to tackle the severity of the backlog we now see. However, the data only goes back to 2007 when the current measure of waiting was introduced (true whether we refer to waiting times or lists). Though it is not straightforward to make comparisons between 2021 and the years before 2007, we can show that waiting times were worse when the NHS began its battle against long waits around the year 2000 and yet it did both drive them down and keep them down until the recent deterioration.

This long read focuses on core waiting times for elective care rather than waiting in A&E or for other specific services such as cancer or mental health.

Waiting times in the 1990s and 2000s

Definitions and measures of waiting times have changed in fundamental ways over the years. However, it is possible to compare these different sets of data and Appendix A provides more detailed data and analysis.

The data reveals that waiting times for patients have been challenging before. For example, as late as 2001, the number of patients waiting more than six months from the time they were first referred by a GP to their first outpatient attendance reached more than 150,000. People who then needed further treatment joined queues for diagnostic tests and for admission to hospital (usually for an operation). In 2001, there were also more than 12,000 patients waiting more than 15 months for a hospital admission (ie, after they had already waited to see a consultant and for any diagnostics to be completed). Waiting times in the 1990s were sometimes even longer – in March 1991 more than 50,000 patients waited more than two years for admission to hospital1  (again, after having waited first for an appointment with a consultant and any diagnostic tests).

Appendix A sets out to compare the current ‘referral-to-treatment’ waiting times (which measure waiting from the point a GP refers a patient until they are admitted for treatment) to these older measures of waiting.

What can we learn from the last drive to bring down waiting times?

The decade after 2000 saw sustained rapid growth in NHS spending and new targets (backed by performance management) set on reducing waiting times. The rise in spending and the use of targets are perhaps the best remembered levers for change in those years, but there is more to learn than the requirement for money and direction.

First, the current monthly release of statistics on hospital waiting times tends to concentrate attention on the acute sector, but the relative lack of routine data in key areas including primary care, community and mental health services doesn’t mean that they aren’t as important as the acute sector. After 2000, alongside the drive to reduce waiting times, came a series of National Service Frameworks to improve quality of care in areas such as cancer, heart disease and mental health, each with their own ‘tsar’ to lead the improvements and these spanned primary and secondary care. Even on waiting times it wasn’t all about hospitals as access to GPs also rose up the political agenda, such that a new waiting times target for general practice was introduced in 2004. By 2008 the agenda moved to focus as much, or more, on quality of care partly because of the success in reducing waits but also to rebalance the NHS towards the other aspects of good-quality care.

'The decisions on priorities post-2021 need to take explicit account of the challenges and opportunities facing the whole health and care system.'

The decisions on priorities post-2021 need to take explicit account of the challenges and opportunities facing the whole health and care system. This becomes critical as increasing elective activity has implications for other services, eg, ring-fencing beds for planned care will reduce the beds available for emergency admissions (other things being equal), and rapid increases in hospital activity will have implications for post-surgical rehabilitation and discharge services.

Second, making sustainable inroads into waiting times takes time. This is because it is not easy or quick to significantly increase NHS (or health care) capacity. One key rate-limiting factor on the ability to increase activity and treat more patients in the 2000s was the availability of staff – and this is still the case. Any plan to reduce waiting times – indeed any NHS plan – needs to build explicitly from an analysis of existing staff and the potential for workforce growth alongside a realistic assessment of any scope for productivity. This is easy to say and hard to do but is essential.

Alongside this there needs to be enough facilities and kit: some diagnostics such as MRI and CT scanners come with a big bill for the initial outlay and take time to procure and install. As many people on the waiting lists are waiting for surgery, there will also need to be sufficient capacity in operating theatres. Both for kit and for staff, improving efficiency can make existing resources go further, but where this is not enough it will take time to build up capacity.

To give an idea of timescale for change, the Labour government elected in 1997 committed to reducing waiting times for NHS treatment. The 18-week target was met in 2008 – a decade later. While quicker progress this time would be desirable, it is important to give careful thought to the trajectory of reducing waiting times. On the one hand, quick fixes that do not increase long-term capacity can be expensive (for example, if they rely on agency staff), and do not deal with any underlying imbalance between demand and supply that led to waiting lists in the first place. On the other hand, raising capacity to both meet new demand and reduce waits risks investing in excess capacity once waiting has been reduced to an acceptable level (and diverts resources from other priorities). Taken together, setting out a plan over a number of years to bring waits back down is a science and an art, and needs analytical support. To state the obvious, this all has to be paid for. The 2000s witnessed high and sustained increases in funding, which may not be available now.

'This time around, careful thought will need to be given to the design of a system that supports multiple objectives: reducing waiting, reducing inequalities, better integration and population health.'

Third, the NHS has already embarked on potentially far-reaching reform to deliver better integration and population health. This includes changes to the financial system and NHS structures. One of the strengths of the 2000s’ approach on waiting was an attempt to align the system behind reducing waits. This included Payment by Results which ensured that those providers that increased activity got paid for it, while reducing waiting times remained a consistent priority for the performance management system over multiple years. This time around, careful thought will need to be given to the design of a system that supports multiple objectives: reducing waiting, reducing inequalities, better integration and population health. While simply re-importing some of the tools used in the 2000s is unlikely to work as they did not support the wider change the NHS is now looking to deliver, it remains the case that if hospitals are to sharply increase activity then their increased costs will need to be paid for.

Fourth, it’s easy to put the waiting times success of the 2000s down to rapid increases in spending and staff. But this ignores the concurrent changes in productivity that also made an important contribution. For example, day-case surgery, rather than keeping patients in for one or more nights, saves money and resources. The proportion of surgery undertaken as day cases rose after 2004 as the system strove to make better use of resources. From the perspective of 2021, the chances of 2000s-style increases in spending and staff are remote and to reduce waiting the balance will need to shift even more from increasing resources to increasing productivity. Recent experience should build confidence about the ability of the NHS to make such a switch: one of the striking positives about the response to Covid-19 was the speed with which the NHS and its partners adopted innovation in order to help manage the crisis. If the circumstances are right the NHS is clearly capable of innovation at scale and pace. One might argue that Covid-19 provided the burning platform for change that was the magic ingredient. Even if this is correct, longer waiting times and aftermath of Covid-19 (including the extent of staff burnout and the need to take action on inequalities) also present just such a burning platform. Seizing the moment on innovation will help the health and care system, and not just in dealing with the backlog of care.

There are, of course, major differences between the situation now and that in the 2000s. For example, the ageing population and the number of people with long-term conditions mean not only that demand for treatment is much higher (as is the capacity of the NHS), but that the balance of support provided in the community (whether health or social care) needs to be greater now both pre- and post-admission. In short, integrated care remains important in the post-Covid-19 recovery.

Finally, while the NHS reduced even longer waits in the 2000s, to do that wasn’t easy then nor will it be now. Last time it took time, money and perseverance. But it is encouraging to remember that the multi-year plans the NHS set out in the 2000s to dramatically reduce waiting times – whether for A&E or elective waits ­– all got delivered within their respective timetables.

Appendix A

Comparisons of NHS waiting times

It is not straightforward to compare waiting times over many years because of the fundamental changes in the way waiting times have been measured, with the current measures coming into use in 2007. Yet the real drive to reduce waiting began with The NHS Plan of 2000, fully seven years before. The main measures of waiting are set out below.

Referral to treatment
The current referral-to-treatment standard for waiting times measures the entire patient pathway from the point at which a GP refers the patient to secondary care to the point the patient is admitted for treatment (whatever that treatment may be). For example, this includes the time taken for a patient to be seen by a hospital consultant after a GP has referred them (the wait for the first outpatient attendance); any diagnostics that may subsequently be requested; and then, should a hospital admission be necessary, the wait until the patient is admitted, or more generally, the wait until ‘treatment’ has begun. The standard for this entire pathway is a maximum of 18 weeks.

Stage of treatment
Before the introduction of the referral-to-treatment standard, the NHS measured waits for individual steps in a patient’s pathway, eg, the waiting time for the first appointment with a hospital consultant; the waiting time for diagnostic tests; and the waiting time for admission to hospital once a consultant decided it was necessary (referred to as ‘the decision to admit’). This stage in the pathway was commonly referred to as ‘the’ waiting list although it was only one of a number of such waits. There was no record of how long it took any single patient to get through this whole process. To add to complexity, the recording of these stages of treatment measures varied over time and the data on waiting for diagnostic tests were only measured from 2006.

It is not possible to simply add up stages-of-treatment waiting times to create a referral-to-treatment waiting time, as no individual patient data was recorded and there was no clear sense of an individual’s patient pathway (eg, how many diagnostic tests – and waits – might have been done or how many outpatient attendances were needed). However, referral-to-treatment waiting times were collected from 2007 and stage-of-treatment waiting times were still recorded into 2010, so there was an overlap between the two sets of measures. This long read uses a simple method to make comparisons across these different waiting times.

  • We compared the latest waiting times statistics (March 2021) to the earliest measure of waiting times calculated on the same basis (ie both based on referral to treatment waiting times). This data set goes back to August 2007. Across the various measures of waiting, 2007 was worse than 2021, though not always by a large margin.

  • We then compared stage-of-treatment data for August 2007 to 2000, looking to see if the experience of waiting in 2000 was better or worse and by how much. If waiting in 2000 was much worse than 2007, which in turn was somewhat worse than 2021, then this tells us that waiting in 2021 – however bad it may be – was not as bad as 2000.

Why pick the year 2000? In 2000, the government published The NHS Plan. This began almost a decade of above average spending growth, a major expansion of the workforce and a consistent drive to reduce waiting times across all aspects of the NHS. This is not to say that waiting was at its worst in 2000 (it was not, as we can show), but it was the point at which the government and the NHS made the strategic decision to try to end the NHS’s historic problem with waiting times.

Waiting times for diagnostics tests were first recorded in 2006 and the data is still collected in 2021. This means we can make a direct comparison between 2006 and 2021, even if 2006 was some time after the rapid increases in NHS expenditure began.

A very simple comparison can also be made between average (median) waiting times. Under the current system (ie, measuring from the point a patient was first referred until they are admitted for treatment), in March 2021 this median waiting time was 11.6 weeks. In September 2000 the only average waiting time that was measured was between the point at which a consultant decided to admit a patient and the point the patient was admitted. This was 13.0 weeks. However, this ignores the wait between GP referral and first outpatients, any subsequent diagnostics or further outpatient attendances. This means that, in 2021, the whole pathway to treatment is shorter, on average, than the old inpatient wait alone in 2000, ignoring all the other steps in the pathway. Making this comparison does, of course, rely on the referral-to-treatment wait remaining lower than the old inpatient wait.

Waiting times in March 2021 and August 2007

Referral-to-treatment waiting times, August 2007 and March 2021

Median (weeks)% within 18 weeksNumber more than 52 weeks% more than 52 weeks
August 200714.357.2578,68213.8
March 202111.664.4436,1278.8

Other measures of referral to treatment (data on admitted unadjusted pathways and non-admitted pathways) show a similar picture between August 2007 and March 2021. The poorer performance on waiting in 2007 is consistent across these measures, even if the difference is not always substantial.

Waiting times in August 2007 and September 2000: first outpatient attendance waiting times

The NHS Plan was published in July 2000, so we have used the first recorded waiting times after that point (September 2000) as the reference.

The recording of waiting times for first outpatient attendance changed between September 2000 and August 2007 (for example, the number of people waiting for shorter waiting times were not recorded in 2000 and average waiting times were not calculated until 2004). Even so, the data tells a consistent story of much longer waits in 2000.

Outpatients not seen (outpatient waiting list), August 2007 and September 2000

+11 weeks +13 weeks (3 months)+17 weeks+26 weeks (6 months)
August 2007 3,66915770-
September 2000Not recorded396,033Not recorded 113,907

At August 2007, the average wait for a first outpatient attendance was around 3.5 weeks (median 3.4 weeks; mean 3.8 weeks). Average waits were not recorded in 2000 and were introduced first in 2004. The poorest performance on average waits for first outpatients were recorded in August and December 2004 when they were around 5.5 to 6 weeks (median 5.5 weeks; mean 6 weeks). It is likely average waits were shorter in 2004 than 2000, as, for example, by 2004 there were no patients waiting more than 6 months for their first outpatient attendance whereas there were 113,907 such waits in September 2000. The total size of the list for first outpatients was not recorded in 2000.

The worst performance on waits of more than 6 months for first outpatient attendance occurred in the first month the data was collected (June 1999), when 157,327 patients had waited this long. The worst performance on waits of more than 3 months came in September 2001 when 497,444 patients had waited this (compared to 157 in 2007).

A line graph showing that the measure of outpatient waits changed several times between September 2000 and August 2007, but all measures showed a reduction in waiting times

Waiting times in August 2007 and September 2000: inpatient waiting times

Once seen by a consultant and after any diagnostic tests, patients could be added to the waiting list for admission to hospital. These waiting times were measured on a more consistent basis than waiting for outpatients and over a longer time frame, beginning in 1993.

Waiting times for inpatient admission, August 2007 and September 2000

Total inpatient listMore than 8 weeksMore than 3 monthsMore than 6 monthsMore than 9 monthsMore than 12 monthsMore than 15 monthsMore than 18 months
August 2007636,612 223,085 79,5343200000
September 2000 1,020,586Not recorded509,000271,819128,29050,269 12,9812

Average waiting times for inpatient admission were 6–7 weeks in August 2007 (5.8 weeks median; 6.8 weeks mean) and between 13 and 19 weeks in September 2000 (13 weeks median; 18.9 weeks mean). Average waits were at their worst when data collection began in June 1993, when they stood at 15.9 weeks (median) and 22.8 weeks (mean). Very long waits were also at their highest at this point, with, for example, 13,080 patients waiting more than 18 months and 123,779 waiting more than 9 months.

A line graph showing that Between September 2000 and August 2007, waiting times for admission to hospital were significantly reduced

Waiting times for diagnostic tests

Waiting times for diagnostic tests are important for two reasons. First, as part of a pathway from GP (or other) referral to first outpatients on to admission to hospital. Second, as an important part of care in their own right, when patients are referred for diagnostics even without any subsequent admission to hospital.

Waiting times for diagnostic tests were not collected until January 2006. At that point, more than half of patients waited more than 6 weeks for their diagnostic test and just under 30 per cent waited for more than 3 months. Because waits for diagnostic tests are also important in their own right irrespective of whether they are part of a pathway leading to admission, they are still collected as a separate ‘stage of treatment’ data set, even if some will also be part of the referral-to-treatment standard. Therefore, we can compare waiting times for diagnostics from 2006 until 2021, although this does miss out the period between 2000 and 2006, when other waiting times were falling.

The data on waiting for diagnostic tests records the total waiting list size, the number of people waiting more than 6 weeks, the number waiting more than 13 weeks and the median. These all show similar trends. From initially high numbers waiting in 2006, all measures of waiting times fell (steeply) through to 2008 and then remained relatively stable – at least by historic standards – until Covid-19 dramatically increased waiting times above their 2006 levels. However, after some recovery through the latter half of 2020, while waiting times remain high by any recent measure, they have fallen back below the levels of 2006. The increases due to Covid-19 in waiting times for diagnostic tests appear to be greater than those for outpatients and inpatients. This may be partly due to the fact that the time period measured for diagnostic tests (6 weeks) is itself relatively short and the initial Covid-19 lockdown itself lasted longer than 6 weeks.

A line graph showing that the proportion of people waiting more than 6 weeks for a diagnostic test fell dramatically and remained at a low level, until the Covid-19 pandemic