Diagnosis and treatment of UTI in the short and long term
A systematic review of the economic evidence relating to the diagnosis and treatment of urinary tract infections was identified as part of a health technology assessment (HTA) on the diagnosis and management of UTI.143 The review identified only one study that met the inclusion criteria for economic evaluations. The review concluded that the findings of this one study were of limited value for the NHS setting as results were not expressed in quality-adjusted life years (QALYs), some parameter estimates were based on data from the USA and may not be valid in the context of the NHS, probabilistic sensitivity analysis for exploring uncertainty was not undertaken and it did not consider the full range of diagnostic strategies available in an NHS setting. However, the economic model identified in the review did provide a basis for developing the subsequent economic model developed specifically for the health technology assessment.
The cost-effectiveness model developed for the health technology assessment attempted to identify the optimal strategy for the diagnosis and treatment of UTI in the NHS to prevent long-term complications believed to be associated with the illness. It should be noted from the outset that this economic model was based on clinical evidence of the risk of developing established renal failure (ERF) that is now considered to be out of date by the Guideline Development Group (GDG) and on specific assumptions which are not accepted as correct by the GDG (see discussion below). Therefore the economic model is described here to provide an account of the economic evidence which has been developed in the UK to date, although the GDG did not consider it to be adequate to support decision making. An update of the economic model to include more recent clinical evidence (specifically on the risk of ERF) and more realistic assumptions was considered for this guideline. However, because it was not possible to estimate the true risks of ERF, it was not possible to produce a more up-to-date economic model at this time.
The economic model developed specifically for the HTA examines the health outcomes and costs for a range of diagnosis and treatment strategies over the long term. Health outcomes in the model are expressed in terms of decrements in a patient’s quality-adjusted life expectancy, with costs estimated from an NHS perspective and discounted in line with NICE technical guidance (6% for costs). Probabilistic sensitivity analysis was undertaken to explore the uncertainty in the values used in the model.
Data used in the HTA economic model
Where possible, data used in the model were taken from the results of meta-analysis of published clinical studies, although in some cases where such evidence was not available, values were estimated from alternative sources. The prevalence of laboratory-proven infection in children presenting with symptoms suggestive of a UTI was estimated at 22.9%, although this number is highly variable between studies, and for the prevalence of vesicoureteric reflux (VUR) in children who have had UTI was estimated at 28.8%. The short-term model also includes estimates for the performance of diagnostic tests, performance of imaging techniques and costs and effects of illness and treatment. In addition, the long-term model includes estimates of recurrence of UTI, the likelihood of pyelonephritic scarring and ERF, and the costs and effects of illness and treatment.
The model considers 79 diagnostic strategies, based on the evidence for diagnostic test performance. The only sequential testing considered is where the second test was considered as a reference standard (laboratory culture for UTI, MCUG for VUR). Options to treat none (with no diagnostic test) and treat all (with no diagnostic test) were considered, as were strategies to treat all children with antibiotics while undertaking alternative imaging tests for VUR, but no tests for UTI. Other combinations include dipstick urine tests followed by laboratory cultures, imaging tests or both.
Health outcomes in the model are presented as decrements in QALYs, from a baseline value of full health, that are the result of the initial UTI as well as the development of ERF after having had a UTI. Costs are taken from published studies with relevant UK costs (for 2003), other published sources such as the British National Formulary or from NHS service providers.
Discussion of the HTA model
The authors presented their results in terms of net monetary benefit (NMB). This allows for the expected costs and benefits (measured in decrements in QALYs) to be compared using a standard unit of measurement, in this case monetary value, when given the maximum willingness-to-pay (WTP) per QALY. The expected NMB is calculated as the expected total number of QALYs generated multiplied by the WTP per QALY; the expected costs of the strategy are then subtracted from this figure. For example, when the WTP for an additional QALY is zero, the NMB is equal to the expected cost subtracted from zero. The strategy with the greatest NMB at a given level of WTP is then identified as the cost-effective strategy.
The results of the model are highly dependent on two factors: the characteristics of the cohort modelled (sex and age) and the WTP of the provider. As the maximum WTP per QALY is not known, sensitivity analysis using estimates ranging from £0 to £50,000 (in increments of £1,000) was undertaken to estimate the cost-effective strategy at each value. Additionally, owing to uncertainty in the value of model parameters, further analysis was undertaken to estimate the likelihood that a given strategy would be the cost-effective strategy. Results are presented in Tables 8.1 and 8.2.
The model consists of a short-term and long-term element. The short-term element considers the optimal diagnostic strategy for UTI and VUR, while the long-term element considers the cost and quality of life decrement for children depending upon down which path of the decision tree they proceed. Assumptions used in the short- and long-term models were discussed at length by the GDG in their deliberations of the evidence. The GDG had specific concerns about the applicability of the results of the model within the context of this guideline. These concerns related to:
- the level of risk assumed in the model for a first-time UTI in childhood leading to ERF
- the level of disutility inherent in the invasive tests included in some diagnostic strategies
- the wider implications of empirical antibiotic treatment, including the disutility from treatment and the development of personal and population resistance to antibiotics.
Therefore, it was the view of the GDG that, due to the use of highly questionable assumptions, the HTA model does not estimate the likely cost-effectiveness of specific management strategies for UTI in children. The GDG, however, recognised the absence of critical data to adequately update this. For these reasons, the diagnosis and treatment strategies recommended in the guideline do not reflect the results of the economic evaluation reported here.
Additional economic analysis undertaken for the guideline
The cost-effectiveness of diagnostic and treatment strategies for UTI focusing on the short term has been evaluated in adults, but not in children. The HTA examined the question of the most appropriate diagnostic and imaging tests for UTI in children.143 This economic analysis assumed a clear relationship between urine infection, VUR and scarring and took a long-term perspective with regard to outcomes. Clinical evidence reported in this guideline suggests that the strength of the link between urine infection in childhood and ERF may be different than previously believed.
Consideration was given by the GDG to the development of a decision-analysis model to examine the cost-effectiveness of the diagnosis and treatment of UTI in children over the short term. A model was proposed for the guideline to examine the period during which the acute symptoms of a UTI may be present; that is 7 days for an uncomplicated infection and 14 days in the case of an upper urinary tract infection (acute pyelonephritis). The model was to give consideration to the following strategies for urine testing and treatment for UTI, either alone or in appropriate combinations:
- no treatment
- empirical treatment
- urine dipstick testing
- microscopic urine analysis for pyuria and/or bacteriuria
- laboratory culture of a urine sample, with or without identification of the organism.
During the development of this model, it was established that much of the data on the clinical effectiveness of the various diagnostic strategies required was either unavailable or of poor quality and that no robust results could be generated. In particular, it was difficult to find key data on the diagnostic accuracy and effectiveness of some urine tests, including microscopy and urine dipsticks, that was stratified by age in a manner that the GDG felt would be clinically useful (for full details on the evidence for urine testing and its shortcomings, refer to Section 4.6). Cost data to be used in the model that was proposed were to be taken from the previous economic analysis described above and are presented in Table 8.3. Cost data would have been adjusted for inflation where appropriate.
During the development of the guideline, it became clear that there would be insufficient data to enable a reliable estimate of the cost-effectiveness of urine testing strategies to be made. There are a number of important but highly uncertain parameters needed to populate the model and any attempt at modelling would produce highly speculative results in the absence of evidence to inform those parameters. In order to arrive at such an estimate, a model such as the one proposed would be required to examine:
- the likelihood that a child presenting with symptoms suggestive of a UTI had a genuine UTI
- the clinical effectiveness of the diagnostic tests in terms of the number of cases of UTI correctly diagnosed
- the incremental cost of each diagnostic strategy
- the incremental effectiveness of each diagnostic strategy expressed as an appropriate outcome, such as QALYs.
Synthesis of this data could then lead to an estimate of the cost-effectiveness of the strategies tested.
Table 8.3 is replicated from the HTA reported above.143 The prices quoted are for 2003. The HTA did not report the costs of a DMSA scan. The cost of a DMSA scan is around £200–300 in 2007 (published private medical practice prices). The cost of urine collection pads and bags was obtained from the NHS Purchasing and Supply Agency in May 2007 and is reported in Table 8.4.
No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.
The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.
RCOG Press, London
National Collaborating Centre for Women's and Children's Health (UK). Urinary Tract Infection in Children: Diagnosis, Treatment and Long-term Management. London: RCOG Press; 2007 Aug. (NICE Clinical Guidelines, No. 54.) 8, Economic evaluation of management of UTI.