CEYLON COLLEGE OF PHYSICIANS MEDICINE UPDATE Vol. 18 No. 1 Compiled by Dr. Henry N. Rajaratnam MD FCCP FRCP (London) (Hon) FRACP (Hon) FSLCGP FACE 1:1 What’s new in TB?. a) Multi drug resistant TB – defined as resistance to at least Rifampicin and INAH are on the increase and considered incurable with conventional regimes. b) HIV increases multi drug resistance, a
Wansbeckchiropractors.co.ukCite this article as: BMJ, doi:10.1136/bmj.38282.607859.AE (published 29 November 2004)
United Kingdom back pain exercise and manipulation (UK BEAM)
randomised trial: cost effectiveness of physical treatments for back
pain in primary care
UK BEAM Trial Team
Full authorship details are given in the accompanying paper (doi: 10.1136/bmj.38282.669225.AE).
poorer outcomes than those randomised to the control groupand encouraged to keep active.5 A Swedish study found no Objective To assess the cost effectiveness of adding spinal
differences in costs or outcomes between physiotherapy and chi- manipulation, exercise classes, or manipulation followed by ropractic manipulation.6 In contrast, a UK trial comparing exercise (“combined treatment”) to “best care” in general private chiropractic and NHS outpatient treatment found that practice for patients consulting with low back pain.
reductions in time off work more than offset the net health serv- Design Stochastic cost utility analysis alongside pragmatic
ice cost incurred by chiropractic.7 To reduce the uncertainty sur- randomised trial with factorial design.
rounding the cost effectiveness of these physical treatments for Setting 181 general practices and 63 community settings for
back pain, we report the economic evaluation of the UK BEAM physical treatments around 14 centres across the United Kingdom.
Participants 1287 (96%) of 1334 trial participants.
Main outcome measures Healthcare costs, quality adjusted life
years (QALYs), and cost per QALY over 12 months.
Results Over one year, mean treatment costs relative to “best
“Best care” in general practice (the “comparator” treatment)—We care” were £195 ($360; €279; 95% credibility interval £85 to trained practice teams in “active management” and provided The £308) for manipulation, £140 (£3 to £278) for exercise, and £125 (£21 to £228) for combined treatment. All three active Exercise programme—This comprised an initial assessment and treatments increased participants’ average QALYs compared up to nine classes in community settings over 12 weeks.10 with best care alone. Each extra QALY that combined Spinal manipulation package—The UK chiropractic, osteo- treatment yielded relative to best care cost £3800; in economic pathic, and physiotherapy professions agreed to use a package of terms it had an “incremental cost effectiveness ratio” of £3800.
techniques developed by a multidisciplinary group, during eight Manipulation alone had a ratio of £8700 relative to combined treatment. If the NHS was prepared to pay at least £10 000 for Combined treatment—Participants received six weeks of each extra QALY (lower than previous recommendations in the manipulation followed by six weeks of exercise. Treatments were United Kingdom), manipulation alone would probably be the otherwise those given to the manipulation only or exercise only best strategy. If manipulation was not available, exercise would have an incremental cost effectiveness ratio of £8300 relative tobest care.
Conclusions Spinal manipulation is a cost effective addition to
We randomised participants between these four interventions.
“best care” for back pain in general practice. Manipulation We also randomised participants receiving manipulation alone probably gives better value for money than manipulation between private and NHS premises. Thus the study had a three by two factorial design. As the accompanying clinical paper didnot find statistically significant differences in outcome between Introduction
manipulation in NHS and private premises, this paper analysesthe simpler two by two factorial trial.
Back pain is a major economic problem. Before the accompany-ing clinical paper,1 little evidence existed for the effectiveness of Data collection
two commonly used treatments—exercise classes and spinal Participants completed questionnaires, including the EQ-5D manipulation. Until the UK BEAM trial, little evidence existed health status instrument,12 13 at baseline, three months, and 12 for the cost effectiveness of those treatments.2 3 months. Over the same period they recorded use of health An economic evaluation in British primary care found that care—including hospital stays, visits to secondary and primary physiotherapy led exercise classes were less expensive and more care, and physical therapists, both private and within the NHS— effective than general practice care alone.4 In contrast, a Finnish whether related to back pain or not. Physical therapists study found that patients randomised to exercise, different from completed record forms, including the number of treatments that investigated by the British study, had higher costs and BMJ Online First bmj.com
So we estimated the mean costs of, and mean QALYs gained Table 1 Reported cost of units of healthcare resource
by, each of four distinct treatments. As most trials compare just Healthcare resource
Cost per unit (£)
two treatments, we adopted a more general approach. Firstly, we ranked treatments by mean cost, starting from the least costly.
Secondly, we calculated incremental cost effectiveness ratios for all treatments by dividing incremental costs by incremental QALYs. Finally, we excluded from the comparison “dominated” Class (assuming mean of 10 participants per class)6 treatments and treatments subject to “extended dominance,”20 and we recalculated ratios if necessary. A treatment is dominated if it generates worse health outcomes and costs more than an alternative treatment. Extended dominance occurs when a treat- ment is less effective and has a higher incremental cost effective- ness ratio than an alternative treatment.20 Trial participants registered with the same general practice formed clusters within centres, and centres formed clusters within the trial. In these circumstances, the use of standard cost utility methods may yield misleading results.21 Hence we used Bayesian Markov Chain Monte Carlo methods within the statis- tical package MLwiN to undertake bivariate multilevel analysis.22 Because evidence about the cost effectiveness of physical treatments for back pain is weak, we started by giving each of the four treatments “uninformative” prior probabilities. In other words, we made no assumptions about the probabilities beforeUK BEAM that one treatment was more effective or cost more Unit costs
To contribute to health policy for an expensive condition, we To report the uncertainty due to sampling variation, we conducted an economic analysis from the perspective of health calculated Bayesian credibility intervals (Bayesian analogue of care. Participants’ follow up periods lay between August 1999 95% confidence intervals) and plotted cost effectiveness accept- and April 2002. We therefore used unit costs in pounds sterling ability curves.23 24 As UK BEAM compared four rather than two at 2000-1 prices to value the resources they used (table 1). We did treatments, we plotted multi-treatment cost effectiveness accept- not adjust or discount the costs, as we focused on effects over ability curves. These curves show the posterior (after UK BEAM) only one year. We costed NHS care from national averages for probability that each strategy is better than the other three across England.14 15 We costed private care by using information from a the range of values that decision makers may pay to achieve an major insurance provider.16 As the accompanying paper found additional QALY. This assumes that these people have no difference in clinical outcome between manipulation in maximum values that they are willing to pay for an additional private and NHS premises,1 our main economic analysis used QALY. It is this “ceiling” against which they should compare esti- costs for the less expensive NHS premises. Nevertheless, we mated incremental cost effectiveness ratios and read the explored this assumption in a sensitivity analysis.
probability that the corresponding treatment is “best.” This deci- Health outcomes
sion oriented formulation selects as “best” or “cost effective” that The EQ-5D measures health on five three point scales—mobility, treatment, with an incremental cost effectiveness ratio below the self care, usual activities, pain-discomfort, and anxiety- ceiling, likely to be more effective than competing treatments.
depression—thus putting participants into one of 243 ( = 35) This is not necessarily the strategy with the lowest ratio, as that health states.12 A large British sample valued these states on a may generate fewer QALYs.23 To cover scenarios in which either “utility” scale on which being dead scores zero and perfect health exercise or manipulation is not available, we also plotted scores one.13 We estimated how many quality adjusted life years two-treatment cost effectiveness acceptability curves to compare (QALYs) participants had experienced over their year in the UK best care with manipulation alone or exercise alone.
BEAM trial by calculating “areas under (health utility) curves.”17 Finally, we did three sensitivity analyses to explore how For example, if they reported that their utility averaged 0.5, we dependent the results were on participants’ estimates of total calculated that they had experienced half a QALY over the year costs and our estimates of unit costs. The first analysis examined in UK BEAM. To avoid bias we adjusted for differences in base- the influence of cost “outliers,” very large healthcare costs reported by a few participants. We excluded those participants Cost utility analysis
whose costs exceeded £2000 ($3700; €2900) (more than eight Because the accompanying clinical paper found statistically times the median cost) in case their chance allocation between significant interactions between manipulation and exercise,1 it compares four distinct treatments within the factorial design.19 The other two sensitivity analyses assessed the influence of Although costs show no interaction between treatments, this the unit costs of manipulation. One did so by costing the paper also compares these four treatments for three reasons.
scenario in which the NHS buys all manipulation from the Firstly, as costs vary much more than clinical outcomes, this is private sector using private manipulation costs. The other did so prudent. Secondly, those people responsible for allocating by costing the scenario in which the NHS buys half its manipula- resources need to choose between these four treatments. Finally, tion from the private sector, using private costs when trial this epitomises the Bayesian statistical approach adopted in this manipulation took place in private premises and NHS costs BMJ Online First bmj.com
Table 2 Reported mean (SD) cost (£) of health care over 12 months by treatment group
Best care in general practice
Best care plus exercise alone
Best care plus manipulation alone
Best care plus manipulation and
*Including spinal manipulation or exercise class outside UK BEAM.
a mean of 0.041 (95% credibility interval 0.016 to 0.066) QALYsper participant, combined treatment generated 0.033 ( − 0.001 to Recruitment
0.067), and exercise generated 0.017 ( − 0.017 to 0.051).
We recruited 1334 participants from 181 practices around 14centres across the United Kingdom. Of these, 1287 (96.5%) Cost utility analysis
yielded enough data for inclusion in the economic analysis; 326 When manipulation and exercise are both available, combined received best care in general practice, 297 received best care plus treatment generates 0.033 more QALYs than does best care at exercise, 342 received best care plus manipulation, and 322 an additional cost of £125, yielding an incremental cost effective- received best care plus combined treatment.
ness ratio of £3800 (table 4). This achievement dominates that of Clinical outcomes
exercise alone, which costs more and achieves less over 12 The accompanying clinical paper reports that exercise achieved months. Manipulation alone, however, can generate 0.008 more a small functional benefit at three months but not at one year; QALYs than combined treatment for an extra £70, yielding a manipulation achieved a small to moderate benefit at three ratio of £8700 relative to combined treatment. If exercise is not months and a small benefit at one year; and combined treatment available, however, manipulation generates 0.041 more QALYs achieved a moderate benefit at three months and a small benefit than best care, yielding an incremental cost effectiveness ratio of at one year (all statistically significant).1 These benefits were spe- £4800. If manipulation is not available, exercise generates 0.017 cific to back pain, in contrast to the general health benefits deter- more QALYs than best care, yielding a ratio of £8300.
The cost effectiveness acceptability curves in the top panel of the figure show the probability that each of the four treatments is better than the other three when all are available. If the ceiling Combined treatment had the highest therapy costs but the low- was only £2000 per QALY, the top panel shows 74% probability est subsequent hospital costs (table 2). So it cost only £125 (95% that best care would be the best strategy. If the ceiling was £5000 credibility interval £21 to £228) more than best care, whereas per QALY, combined treatment has a lower incremental cost exercise cost £140 (£3 to £278) more than best care, and effectiveness ratio than this; the top panel of the figure shows a manipulation cost £195 (£85 to £308) more.
46% chance that it would be best. If the ceiling was £15 000 per Health outcomes
QALY (lower than implied by previous recommendations by the Physical interventions improved EQ-5D scores more than did National Institute for Clinical Excellence25), manipulation alone best care (table 3). Relative to best care, manipulation generated has a lower incremental cost effectiveness ratio than this; the top Table 3 Utilities and QALYs over 12 months by treatment group
Best care in general
Best care plus exercise
Best care plus manipulation
Best care plus manipulation
and exercise (n=322)
Difference (95% credibility interval) in mean QALYs relative to best care in general practice* QALY=quality adjusted life year.
*Estimated by analysis of covariance with adjustment for baseline EQ-5D score and then rounded to three significant figures.
BMJ Online First bmj.com
Table 4 Cost utility analysis by scenario and treatment group*
effectiveness ratio to
nearest £100 or comment
Manipulation and exercise both available (n=1287)
Exercise not available (n=623)
Manipulation not available (n=668)
*Markov Chain Monte Carlo estimates from bivariate multilevel model.
†Estimated by analysis of covariance with adjustment for baseline EQ-5D score and then rounded to three significant figures.
panel shows a 50% probability that it would be best. The costeffectiveness acceptability curve in the middle panel of the figure shows the probability that manipulation is better than best care when exercise is not available; and vice versa for the curve in thebottom panel.
To assess the robustness of these results to the presence of “out- liers,” we excluded the 51 participants (9, 16, 16, and 10 from best care, exercise alone, manipulation alone, and combined treatment respectively) whose healthcare costs exceeded £2000.
Manipulation achieves extended dominance over both exercise and combined treatment (table 5). It is thus the only alternativeto best care, with an incremental cost effectiveness ratio of £3000 per additional QALY. At a ceiling of £10 000 per QALY, Cost effectiveness acceptability curves. Top: manipulation and exercise available; manipulation has a 73% chance of being best. If manipulation middle: exercise not available; bottom: manipulation not available.
alone were not available, exercise would have a ratio of £4100.
ICER=incremental cost effectiveness ratio The second sensitivity analysis used private costs for manipu- lation that took place in private premises. Combined treatment Strengths and weaknesses of the study
now achieves extended dominance over exercise, with an Although two of our three sensitivity analyses—those that used incremental cost effectiveness ratio of £6600 compared with best larger unit costs in whole or in part—changed these critical care (table 5). Manipulation alone then has a ratio of £8700 rela- thresholds a little, they did not alter the essentials of these conclusions. The other sensitivity analysis, which removed 51 The third sensitivity analysis used private unit costs for all “outliers” from the UK BEAM dataset, was more favourable to manipulation within the trial. The findings are analogous to manipulation than was the primary analysis. Under this scenario those in the second scenario. Exercise is subject to extended manipulation cost only £3000 per QALY relative to best care in dominance, and combined treatment has an incremental cost effectiveness ratio of £8600 compared with best care (table 5).
We believe that this is the first study of physical therapy for Manipulation alone then has a ratio of £10 600 relative to com- low back pain to show convincingly that both manipulation alone and manipulation followed by exercise provide cost effec-tive additions to care in general practice. Indeed, as we trained Discussion
practice teams in the best care of back pain, we may have under-estimated the benefit of physical therapy when compared with Principal findings
“usual care” in general practice. The detailed clinical outcomes This economic evaluation supports and extends the findings of reported in the accompanying paper reinforce these findings by the clinical evaluation of the UK BEAM trial reported in the showing that the improvements in health status reported here accompanying paper.1 If decision makers value additional quality reflect statistically significant improvements in function, pain, adjusted life years (QALYs) at much less than £3800, “best care” disability, physical and mental aspects of quality of life, and in general practice is probably the best strategy. If their valuation lies between £3800 and £8700, spinal manipulation followed byexercise classes (“combined treatment”) is likely to be the best Unanswered questions
treatment. If their valuation is well above £8700, manipulation Funding constraints prevented us from following up participants alone is probably the best treatment.
for more than 12 months. Given that they continued to show BMJ Online First bmj.com
Table 5 Sensitivity analyses by treatment group*
What is already known on this topic
effectiveness ratio to
nearest £100 or comment
Excluding participants with health care costing more than £2000 over 12 months (n=1266)
Little evidence exists for the effectiveness and cost effectiveness of two commonly used treatments—exercise Subject to extendeddominance by manipulation What this study adds
Spinal manipulation, exercise classes, and manipulation followed by exercise all increased participants’ quality of life NHS provides manipulation for only 50% of patients (n=1287)
over 12 months by more than did “best care” in general Adding spinal manipulation to best care in general practice is effective and cost effective for patients in the United If the NHS can afford at least £10 000 for each quality adjusted life year yielded by physical treatments, Manipulation provided only in private premises (n=1287)
manipulation alone probably gives better value for money Subject to extendeddominance by manipulationand exercise Meaning of the study
Adding spinal manipulation to best care in general practice is effective and cost effective for patients with back pain in the £10 600 relative tomanipulation and exercise United Kingdom. If the NHS can afford more than £10 000 foran extra QALY, manipulation alone probably gives better value QALY=quality adjusted life year.
*Markov Chain Monte Carlo estimates from bivariate hierarchical multilevel model.
for money than manipulation followed by exercise. These †Estimated by analysis of covariance with adjustment for baseline EQ-5D score and then conclusions hold even if the NHS has to buy spinal manipulation rounded to three significant figures.
We thank all participants—patients, primary care staff, and collaborators benefits of treatment at 12 months, the cost effectiveness of both listed in the accompanying clinical paper—for their contributions. We thank manipulation and combined treatment may be better than we Mark Sculpher and Daphne Russell for advice on analysis. Members of the UK BEAM Trial Team: Ian Russell, Martin Underwood, Stephen Brealey,Kim Burton, Simon Coulton, Amanda Farrin, Andrew Garratt, Emma Har- Commissioning decisions should depend on local circum- vey, Louise Letley, Andrea Manca, Jeannett Martin, Jennifer Klaber Moffett, stances, notably the availability of spinal manipulation and exer- Veronica Morton, David Torgerson, Madge Vickers, Ken Whyte, Melanie cise physiotherapists. Although combined therapy is an attractive option, this depends on an ample supply of both trained Contributors: See accompanying clinical paper.
manipulators prepared to work for the NHS and exercise physi- Funding: Medical Research Council (research costs); NHS in England, otherapists with access to suitable premises. As back pain is a Northern Ireland, Scotland, and Wales (excess treatment and servicesupport costs).
common problem, making manipulation generally available will Competing interests: LL, JM, MU, MV, and KW have received salaries from require many therapists. In the United Kingdom there are 2100 the MRC. MU has received fees for speaking from Menarini Pharmaceuti- registered chiropractors, 3200 registered osteopaths, and about cals, the manufacturers of dexketoprofen and ketoprofen, and Pfizer, the 5000 manipulative physiotherapists (Ann Thomson, Chair of manufacturers of celecoxib and valdecoxib. The other 12 authors have British Association of Chartered Physiotherapists in Manipula- tion, on behalf of the Chartered Society of Physiotherapy, Ethical approval: The Northern and Yorkshire multicentre research ethicscommittee and 41 local research ethics committees approved the trial pro- personal communication, 2003). According to the unit costs we used in our analysis, they can achieve higher incomes in privatepractice than in the NHS. In the short term it may be difficult to UK BEAM Trial Team. United Kingdom back pain exercise and manipulation (UK make manipulative or combined treatment generally available BEAM) randomised trial: effectiveness of physical treatments for back pain in primarycare. BMJ 2004;329:doi = 10.1136/bmj.38282.669225.AE.
Baldwin ML, Côté P, Frank JW, Johnson WG. Cost-effectiveness studies of medical and Whereas physiotherapists can rapidly train to deliver the chiropractic care for occupational low back pain: a critical review of the literature. SpineJ 2001;1:138-47.
exercise package, insufficient trained manipulators are available Maetzel A, Li L. The economic burden of low back pain: a review of studies published in the United Kingdom to meet potential demand, and it will between 1996 and 2001. Best Pract Res Clin Rheumatol 2002;16:23-30.
take several years to produce additional manipulators. Indeed, if Klaber Moffett J, Torgerson D, Bell-Syer S, Jackson D, Llewlyn-Phillips H, Farrin A, et al.
Randomised controlled trial of exercise for low back pain: clinical outcomes, costs, and this needs new training programmes, it may be decades before preferences. BMJ 1999;319:279-83.
the NHS can implement these findings. Fortunately, using Malmivaara A, Hakkinen U, Heinrichs M, Koskenniemi L, Kuosma E, Lappi S, et al. Thetreatment of acute low back pain—bed rest, exercises, or ordinary activity? N Engl J Med private costs for manipulation had little effect on the choice of best treatment. Purchasing manipulation from the private sector Skargren EI, Carlsson PG, Oberg BE. One-year follow-up comparison of the cost andeffectiveness of chiropractic and physiotherapy as primary management for back pain.
to provide treatment within the NHS would still represent good value for money if decision makers were willing to pay £10 000 Meade TW, Dyer S, Browne W, Townsend J, Frank AO. Low back pain of mechanicalorigin: randomised comparison of chiropractic and hospital outpatient treatment. BMJ BMJ Online First bmj.com
Underwood M, O’Meara S, Harvey E, UK BEAM Trial Team. The acceptability to pri- 21 Manca A, Rice N, Sculpher MJ, Briggs AH. Assessing generalisability by location in mary care staff of a multidisciplinary training package on acute back pain guidelines.
trial-based cost-effectiveness analysis: the use of multilevel models. Health Econ 2004 Jun 14 [Epublication ahead of print].
Roland M, Waddell G, Klaber Moffett J, Burton AK, Main CJ, Cantrell T. The back book.
22 Rasbash J, Browne W, Goldstein H. A user’s guide to MLwiN Version 2.1. London: Insti- tute of Education, University of London, 2000.
10 Klaber Moffett J, Frost H, UK BEAM Trial Team. Back to fitness programme: the 23 Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of manual for physiotherapists to set up the classes. Physiotherapy 2000;85:295-305.
cost-effectiveness acceptability curves. Health Econ 2001;10:779-87.
11 Harvey E, Burton AK, Moffett JK, Breen A, UK BEAM Trial Team. Spinal manipulation 24 Fenwick E, O’Brien BJ, Briggs AH. Cost-effectiveness acceptability curves—facts, for low-back pain: a treatment package agreed to by the UK chiropractic, osteopathy fallacies and frequently asked questions. Health Econ 2004;13:405-15.
25 Raftery J. NICE: faster access to modern treatments? Analysis of guidance on health and physiotherapy professional associations. Man Ther 2003;8:46-51.
technologies. BMJ 2001;323:1300-3.
12 Kind P. The EuroQoL instrument: an index of health-related quality of life. In: Spilker B, ed. Quality of life and pharmacoeconomics in clinical trials. Philadelphia: 13 Kind P, Hardman G, Macran S. UK population norms for EQ-5D. York: Centre for Health Correspondence to: Andrea Manca, research fellow, Centre for Health Economics, Economics, University of York, 1999. (Discussion paper 172.) University of York, York YO10 5DD firstname.lastname@example.org 14 Chartered Institute of Public Finance and Accountancy. The health service database 2002.
15 Netten A, Dennett J, Knight J. Unit costs of health and social care. Canterbury: Personal Social Services Research Unit, University of Kent, 2002.
16 BUPA Hospitals UK. How much will it cost? www.bupahospitals.co.uk/asp/paying/ Amendment
priceguides.asp (accessed 17 Nov 2004).
17 Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements This is Version 2 of the paper. In this version, the text relating in medical research. BMJ 1990;300:230-5.
to the figure has been amended to state that the curve in the 18 Manca A, Hawkins N, Sculpher MJ. Estimating mean QALYs in trial-based cost-effectiveness analysis: the importance of controlling for baseline utility. Health Econ middle panel shows the probability that manipulation is 2004 Oct 20 [Epublication ahead of print].
better than best care when exercise is not available, and vice 19 McAlister FA, Straus SE, Sackett DL, Altman DG. Analysis and reporting of factorial versa for the curve in the bottom panel [these were trials: a systematic review. JAMA 2003;289:2545-53.
incorrectly given the other way round in the previous 20 Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health and medicine.
New York: Oxford University Press, 1996.
BMJ Online First bmj.com
Substance Abuse Treatment Breaking News for the T reatment Field Acamprosate: A New Medication for Alcohol Use Disorders What is acamprosate? glutamate and gamma-aminobutyric acid (GABA) neurotransmitter systems. Although acamprosate’s Acamprosate (calcium acetyl homotaurinate) is a mechanism of action has not been clearly established, new prescription medication t