Variations in NHS admissions at a glance

Variations in admissions to NHS hospitals are the source of a great deal of consternation. Over the long-run, admissions and the volume of activity required of the NHS have increased, without equivalent increases in funding or productivity. Over the course of the year, there are repeated claims of crises as hospitals are ill-equipped for the increase in demand in the winter. While different patterns of admissions at weekends relative to weekdays may be the foundation of the ‘weekend effect’ as we recently demonstrated. And yet all these different sources of variation produce a singular time series of numbers of daily admissions. But, each of the different sources of variation are important for different planning and research aims. So let’s decompose the daily number of admissions into its various components.


Daily number of emergency admissions to NHS hospitals between April 2007 and March 2015 from Hospital Episode Statistics.


A similar analysis was first conducted on variations in the number of births by day of the year. A full description of the model can be found in Chapter 21 of the textbook Bayesian Data Analysis (indeed the model is shown on the front cover!). The model is a sum of Gaussian processes, each one modelling a different aspect of the data, such as the long-run trend or weekly periodic variation. We have previously used Gaussian processes in a geostatistical model on this blog. Gaussian processes are a flexible class of models for which any finite dimensional marginal distribution is Gaussian. Different covariance functions can be specified for different models, such as the aforementioned periodic or long-run trends. The model was run using the software GPstuff in Octave (basically an open-source version of Matlab) and we have modified code from the GPstuff website.



The four panels of the figure reveal to us things we may claim to already know. Emergency admissions have been increasing over time and were about 15% higher in 2015 than in 2007 (top panel). The second panel shows us the day of the week effects: there are about 20% fewer admissions on a Saturday or Sunday than on a weekday. The third panel shows a decrease in summer and increase in winter as we often see reported, although perhaps not quite as large as we might have expected. And finally the bottom panel shows the effects of different days of the year. We should note that the large dip at the end of March/beginning of April is an artifact of coding at the end of the financial year in HES and not an actual drop in admissions. But, we do see expected drops for public holidays such as Christmas and the August bank holiday.

While none of this is unexpected it does show that there’s a lot going on underneath the aggregate data. Perhaps the most alarming aspect of the data is the long run increase in emergency admissions when we compare it to the (lack of) change in funding or productivity. It suggests that hospitals will often be running at capacity so other variation, such as over winter, may lead to an excess capacity problem. We might also speculate on other possible ‘weekend effects’, such as admission on a bank holiday.

As a final thought, the method used to model the data is an excellent way of modelling data with an unknown structure without posing assumptions such as linearity that might be too strong. Hence their use in geostatistics. They are widely used in machine learning and artificial intelligence as well. We often encounter data with unknown and potentially complicated structures in health care and public health research so hopefully this will serve as a good advert for some new methods. See this book, or the one referenced in the methods section, for an in depth look.


Alastair Canaway’s journal round-up for 31st October 2016

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

Ethical hurdles in the prioritization of oncology care. Applied Health Economics and Health Policy [PubMedPublished 21st October 2016

Recently between health economists, there has been significant scrutiny and disquiet directed towards the Cancer Drugs Fund with Professor Karl Claxton describing it as “an appalling, unfair use of NHS resources”. With the latest reorganization of the Cancer Drugs Fund in mind, this article examining the ethical issues surrounding prioritisation of cancer care was of particular interest. As all health economists will tell you, there is an opportunity cost with any allocation of scarce resources. Likewise, with prioritisation of specific disease groups, there may be equity issues with specific patients’ lives essentially being valued more greatly than those suffering other conditions. This article conducts a systematic review of the oncology literature to examine the ethical issues surrounding inequity in healthcare. The review found that public and political attention often focuses on ‘availability’ of pharmacological treatment in addition to factors that lead to good outcomes. The public and political focus on availability can have perverse consequences as highlighted by the Cancer Drugs Fund: resources are diverted towards availability and away from other more cost-effective areas, and in turn this may have had a detrimental effect on care for non-cancer patients. Additionally, by approving high cost, less cost-effective agents, strain will be placed upon health budgets and causing problems for existing cost-effectiveness thresholds. If prioritisation for cancer drugs is to be pursued then the authors suggest that the question of how to fund new therapies equitably will need to be addressed. Although the above issues will not be new to most, the paper is still worth reading as it: i) gives an overview of the different prioritisation frameworks used across Europe, ii) provides several suggestions for how, if prioritization is to be pursued, it can be done in a fairer manner rather than simply overriding typical HTA decision processes, iii) considers the potential legal consequences of prioritisation and iv) the impact of prioritisation on the sustainability of healthcare funding.

Doctor-patient differences in risk and time preferences: a field experiment. Journal of Health Economics Published 19th October 2016

The patient-doctor agency interaction, and associated issues due to asymmetrical information is something that was discussed often during my health economics MSc, but rarely during my day to day work. Despite being very familiar with supplier induced demand, differences in risk and time preferences in the patient-doctor dyad wasn’t something I’d considered in recent times. Upon reading, immediately, it is clear that if risk and time preferences do differ, then what is seen as the optimal treatment for the patient may be very different to that of the doctor. This may lead to poorer adherence to treatments and worse outcomes. This paper sought to investigate whether patients and their doctors had similar time and risk preferences using a framed field experiment with 300 patients and 67 doctors in Athens, Greece in a natural clinical setting. The authors claim to be the first to attempt this, and have three main findings: i) there were significant time preference differences between the patients and doctors – doctors discounted future health gains and financial outcomes less heavily than patients; ii) there were no significant differences in risk preferences for health with both doctors and patients being mildly risk averse; iii) there were however risk preference differences for financial impacts with doctors being more risk averse than patients. The implication of this paper is that there is potential for improvements in doctor-patient communication for treatments, and as agents for patients, doctors should attempts to gauge their patient’s preferences and attitudes before recommending treatment. For those who heavily discount the future it may be preferable to provide care that increases the short term benefits.

Hospital productivity growth in the English NHS 2008/09 to 2013/14 [PDF]. Centre for Health Economics Research Paper [RePEcPublished 21st October 2016

Although this is technically a ‘journal round-up’, this week I’ve chosen to include the latest CHE report as I think it is something which may be of wider interest to the AHEBlog community. Given limited resources, there is an unerring call for both productivity and efficiency gains within the NHS. The CHE report examines the extent to which NHS hospitals have improved productivity: have they made better use of their resources by increasing the number of patients they treat and the services they deliver for the same or fewer inputs. To assess productivity, the report uses established methods: Total Factor Productivity (TFP) which is the ratio of all outputs to all inputs. Growth in TFP is seen as being key to improving patient care with limited resources. The primary report finding was that TFP growth at the trust level exhibits ‘extraordinary volatility’. For example one year there maybe TFP growth followed by negative growth the next year, and then positive growth. The authors assert that much of the TFP growth measured is in fact implausible, and much of the changes are driven largely by nominal effects alongside some real changes. These nominal effects may be data entry errors or changes in accounting practices and data recording processes which results in changes to the timing of the recording of outputs and inputs. This is an important finding for research assessing productivity growth within the NHS. The TFP approach is an established methodology, yet as this research demonstrates, such methods do not provide credible measures of productivity at the hospital level. If hospital level productivity growth is to be measured credibly, then a new methodology will be required.


Alastair Canaway’s journal round-up for 12th September 2016

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

Question order sensitivity of subjective well-being measures: focus on life satisfaction, self-rated health, and subjective life expectancy in survey instruments. Quality of Life Research [PubMed] Published 30th April 2016

It’s interesting to see an ‘old’ and well known issue rearing its head within the health economics literature. In this case, the focus is on ordering bias within wellbeing questionnaires. It is established within the psychometric and psychological literature that the location of a question within a survey can influence how respondents interpret the meaning of the question, and therefore their answers. This study sought to empirically examine how ordering in subjective well-being measures (life satisfaction, self-rated health, and subjective life expectancy) affected answers. Given ordering bias is an established concept, it wasn’t too surprising to find notable ordering bias depending on how the questionnaire was ordered. For example, as hypothesised by the authors, placing self-rated health immediately before life satisfaction within the survey led to different values compared to when placed apart. For well-being research, the paper has important implications, particularly in how to best order questionnaires to reduce the impact of prior questions on answers, e.g. keeping self-rated health and life satisfaction questions apart to encourage respondents to independently evaluate each question. Ordering bias is one of those issues that most researchers are aware of, but tend to forget about. As much as anything I feel this is for pragmatic reasons, for example, in terms of ease of producing case report forms and also for facilitating data entry within trials. Ideally, we probably should be randomising the order of questionnaires, whether we can persuade wider trial teams that this is necessary remains to be seen.

You sneeze, you lose: The impact of pollen exposure on cognitive performance during high-stakes high school exams. Journal of Health Economics [PubMed] [RePEcPublished September 2016

As a ‘summer sneezer’ and someone with poor exam results in year 9, it was of great interest to read this article. It is known that health and productivity are intrinsically linked, indeed productivity costs related to health are commonly discussed within health economics circles. Elsewhere there are studies that have identified pollution levels as having significant effects on labour productivity and supply. As any fellow hay fever (seasonal allergic rhinitis) sufferers will attest, hay fever has a direct negative impact on wellbeing. Hay fever is relatively prevalent with over one in five people being reported to suffer (in the Norwegian setting at least). This study combined a large administrative dataset from the Norwegian high school system with daily pollen counts from measurement stations across Norway. Student exam data were matched with location of exams and the pollen count for the area in which the exam took place. Fixed effect panel data methods were used to analyse the data. The primary result found that one standard deviation increase in pollen levels led to a decrease in a student’s exam score by about 2.5% of a standard deviation, the implication of this is that for allergic students, this negative effect is approximately 10% of a standard deviation. This is a notable margin. The paper has an interesting discussion on the potential long term impact of hay fever on allergic students, and their future prospects e.g. impact on university enrolment. To avoid such impacts the paper emphasises the need to diagnose early and optimise treatment for hay fever in children. One final point (and word of caution) would be that the methods don’t prove causality, however as a hay fever sufferer, it was very interesting nonetheless to consider how the condition may have impacted upon my own performance at school.

The fatter are happier in Indonesia. Quality of Life Research [PubMed] Published 31st August 2016

An eye-catching title. In developed countries, being overweight and obese typically has negative connotations, and studies repeatedly suggest this is the case: those who are overweight are less happy. In developing countries however, this is not necessarily true. The paper offers the following reason for this: wealth and obesity are positively correlated in such countries, and likewise, happiness and wealth are positively related. Those who are poor in developing countries literally cannot afford to be obese. In contrast, in developed countries, even lower socioeconomic classes can afford to be obese (and obesity is indeed more prevalent in these classes). With this in mind, this study sought to determine how obesity and happiness were related in Indonesia. The study used a large long term survey of over 22,000 participants over a long time period. As hypothesised, the study found there to be a positive association between obesity and self-reported happiness within Indonesia. The paper in a roundabout way suggests that a different approach to evaluating obesity prevention is required in the developing world. I’m not sure this is necessarily the case, in my experience it is rare to assess obesity prevention interventions with respect to ‘happiness’. It takes me back to a previous journal round-up discussing the maximand within economic evaluation. Obesity, if not immediately, eventually is associated with poor health, therefore there is nothing to suggest that an evaluative framework that seeks to maximise health over happiness will not be sufficient. There are many issues related to the long term evaluation of obesity prevention interventions, particularly those focussed in children (as outlined here), however I think the case stated in this paper is a bit of red herring.

Photo credit: Antony Theobald (CC BY-NC-ND 2.0)