Number needed to treat: caveat emptor

Spinal manipulative therapy for chronic low-back pain: an update of a Cochrane review

STUDY DESIGN

Systematic review of interventions.

OBJECTIVE

To assess the effects of spinal manipulative therapy (SMT) for chronic low-back pain.

SUMMARY OF BACKGROUND DATA

SMT is one of the many therapies for the treatment of low-back pain, which is a worldwide, extensively practiced intervention.

METHODS

Search methods. An experienced librarian searched for randomized controlled trials (RCTs) in multiple databases up to June 2009. Selection criteria. RCTs that examined manipulation or mobilization in adults with chronic low-back pain were included. The primary outcomes were pain, functional status, and perceived recovery. Secondary outcomes were return-to-work and quality of life. Data collection and analysis. Two authors independently conducted the study selection, risk of bias assessment, and data extraction. GRADE was used to assess the quality of the evidence.

RESULTS

We included 26 RCTs (total participants = 6070), 9 of which had a low risk of bias. Approximately two-thirds of the included studies (N = 18) were not evaluated in the previous review. There is a high-quality evidence that SMT has a small, significant, but not clinically relevant, short-term effect on pain relief (mean difference -4.16, 95% confidence interval -6.97 to -1.36) and functional status (standardized mean difference -0.22, 95% confidence interval -0.36 to -0.07) in comparison with other interventions. There is varying quality of evidence that SMT has a significant short-term effect on pain relief and functional status when added to another intervention. There is a very low-quality evidence that SMT is not more effective than inert interventions or sham SMT for short-term pain relief or functional status. Data were particularly sparse for recovery, return-to-work, quality of life, and costs of care. No serious complications were observed with SMT.

CONCLUSIONS

High-quality evidence suggests that there is no clinically relevant difference between SMT and other interventions for reducing pain and improving function in patients with chronic low-back pain. Determining cost-effectiveness of care has high priority.

Spinal manipulative therapy for chronic low-back pain

BACKGROUND

Many therapies exist for the treatment of low-back pain including spinal manipulative therapy (SMT), which is a worldwide, extensively practiced intervention.

OBJECTIVES

To assess the effects of SMT for chronic low-back pain.

SEARCH STRATEGY

An updated search was conducted by an experienced librarian to June 2009 for randomised controlled trials (RCTs) in CENTRAL (The Cochrane Library 2009, issue 2), MEDLINE, EMBASE, CINAHL, PEDro, and the Index to Chiropractic Literature.  

SELECTION CRITERIA

RCTs which examined the effectiveness of spinal manipulation or mobilisation in adults with chronic low-back pain were included. No restrictions were placed on the setting or type of pain; studies which exclusively examined sciatica were excluded. The primary outcomes were pain, functional status and perceived recovery. Secondary outcomes were return-to-work and quality of life.

DATA COLLECTION AND ANALYSIS

Two review authors independently conducted the study selection, risk of bias assessment and data extraction. GRADE was used to assess the quality of the evidence. Sensitivity analyses and investigation of heterogeneity were performed, where possible, for the meta-analyses.

MAIN RESULTS

We included 26 RCTs (total participants = 6070), nine of which had a low risk of bias. Approximately two-thirds of the included studies (N = 18) were not evaluated in the previous review. In general, there is high quality evidence that SMT has a small, statistically significant but not clinically relevant, short-term effect on pain relief (MD: -4.16, 95% CI -6.97 to -1.36) and functional status (SMD: -0.22, 95% CI -0.36 to -0.07) compared to other interventions. Sensitivity analyses confirmed the robustness of these findings. There is varying quality of evidence (ranging from low to high) that SMT has a statistically significant short-term effect on pain relief and functional status when added to another intervention. There is very low quality evidence that SMT is not statistically significantly more effective than inert interventions or sham SMT for short-term pain relief or functional status. Data were particularly sparse for recovery, return-to-work, quality of life, and costs of care. No serious complications were observed with SMT.

AUTHORS' CONCLUSIONS

High quality evidence suggests that there is no clinically relevant difference between SMT and other interventions for reducing pain and improving function in patients with chronic low-back pain. Determining cost-effectiveness of care has high priority. Further research is likely to have an important impact on our confidence in the estimate of effect in relation to inert interventions and sham SMT, and data related to recovery.

Near-Normalization of Glucose and Microvascular Diabetes Complications: Data from ACCORD and ADVANCE

OBJECTIVE

To compare results of clinical trials that assessed the impact of near-normalization of glucose on microvascular complications in type 2 diabetes.

METHODS

ACCORD (N = 10,234) and ADVANCE (N = 11,140) tested the hypothesis that near-normalization of glucose reduces microvascular complications in adults with established type 2 diabetes. Differences in incidence rates (intensive versus standard glucose control) for specific microvascular complications are expressed as 'number needed to treat' (NNT) to prevent one microvascular complication. The impact of blood pressure (BP) control and fenofibrate use on microvascular complications was also assessed.

RESULTS

In ADVANCE, near-normalization of glucose reduced new or worsening nephropathy (NNT = 77 for 5 years to prevent one occurrence), but not eye or foot complications. In ACCORD, near-normalization of glucose did not reduce prespecified composite measures of advanced microvascular complications, and impact on secondary microvascular outcomes was mixed. The ancillary ACCORD Eye Study found reduced progression in retinopathy with near-normalization of glucose (NNT = 32 for 4 years), and with blinded fenofibrate therapy (NNT = 27 for 4 years), but neither intervention reduced vision loss. ADVANCE showed a benefit of intensive BP control (mean BP 133/70 mmHg) on microvascular complications, independent of glucose control.

CONCLUSIONS

End-stage microvascular complications were not altered by near-normalization of glucose. Some early manifestations of microvascular complications were reduced, with inconsistencies across studies in which were affected. These early and inconsistent micro-vascular effects must be weighed against significantly increased severe hypoglycemia, weight gain, and (in ACCORD) increased total mortality (NNT = 94 for 3.5 years for one excess death) consistently found in all prespecified patient subgroups. Alternative clinical strategies, such as moderate BP control or fenofibrate treatment may reduce microvascular complications independent of glucose control. The data strongly support personalized glucose control goals based on clinical factors and patient preferences for outcomes.

The Application of Number Needed to Treat (NNT) to Clinical Problems in Radiotherapy

The goals of this report are: 1) to review the number needed to treat (NNT) concept, which, although well established in many sectors of medicine, is still relatively new to the radiotherapy community; 2) to discuss several clinical radiotherapy examples illustrating the inherent advantages of the NNT approach; and 3) to discuss potential future roles of the NNT concept within radiotherapy.

Estimating the number needed to treat from continuous outcomes in randomised controlled trials: methodological challenges and worked example using data from the UK Back Pain Exercise and Manipulation (BEAM) trial

Background

Reporting numbers needed to treat (NNT) improves interpretability of trial results. It is unusual that continuous outcomes are converted to numbers of individual responders to treatment (i.e., those who reach a particular threshold of change); and deteriorations prevented are only rarely considered. We consider how numbers needed to treat can be derived from continuous outcomes; illustrated with a worked example showing the methods and challenges.

Methods

We used data from the UK BEAM trial (n = 1, 334) of physical treatments for back pain; originally reported as showing, at best, small to moderate benefits. Participants were randomised to receive 'best care' in general practice, the comparator treatment, or one of three manual and/or exercise treatments: 'best care' plus manipulation, exercise, or manipulation followed by exercise. We used established consensus thresholds for improvement in Roland-Morris disability questionnaire scores at three and twelve months to derive NNTs for improvements and for benefits (improvements gained+deteriorations prevented).

Results

At three months, NNT estimates ranged from 5.1 (95% CI 3.4 to 10.7) to 9.0 (5.0 to 45.5) for exercise, 5.0 (3.4 to 9.8) to 5.4 (3.8 to 9.9) for manipulation, and 3.3 (2.5 to 4.9) to 4.8 (3.5 to 7.8) for manipulation followed by exercise. Corresponding between-group mean differences in the Roland-Morris disability questionnaire were 1.6 (0.8 to 2.3), 1.4 (0.6 to 2.1), and 1.9 (1.2 to 2.6) points.

Conclusion

In contrast to small mean differences originally reported, NNTs were small and could be attractive to clinicians, patients, and purchasers. NNTs can aid the interpretation of results of trials using continuous outcomes. Where possible, these should be reported alongside mean differences. Challenges remain in calculating NNTs for some continuous outcomes.

Trial Registration

UK BEAM trial registration: ISRCTN32683578.

Approaching clinical problems in prostate cancer radiotherapy using the number needed to treat (NNT) technique

The goals of this article are to review the application of the number needed to treat (NNT) concept to selected clinical problems in prostate cancer radiotherapy. Particular emphasis will be placed on (1) comparison of radiotherapy with other treatment options for early-stage disease, (2) the role of hormone therapy in addition to radiotherapy over a spectrum of disease presentation, and (3) systematic comparison of adjuvant versus salvage radiotherapy in the post-prostatectomy setting. Limitations of NNT calculations based on non-randomized comparisons also are discussed.

Using economic analyses for local priority setting : the population cost-impact approach

INTRODUCTION

Standard methods of economic analysis may not be suitable for local decision making that is specific to a particular population.

BACKGROUND

We describe a new three-step methodology, termed 'population cost-impact analysis', which provides a population perspective to the costs and benefits of alternative interventions. The first two steps involve calculating the population impact and the costs of the proposed interventions relevant to local conditions. This involves the calculation of population impact measures (which have been previously described but are not currently used extensively) - measures of absolute risk and risk reduction, applied to a population denominator. In step three, preferences of policy-makers are obtained. This is in contrast to the QALY approach in which quality weights are obtained as a part of the measurement of benefit.

METHODS

We applied the population cost-impact analysis method to a comparison of two interventions - increasing the use of beta-adrenoceptor antagonists (beta-blockers) and smoking cessation - after myocardial infarction in a scaled-back notional local population of 100,000 people in England. Twenty-two public health professionals were asked via a questionnaire to rank the order in which they would implement four interventions. They were given information on both population cost impact and QALYs for each intervention.

RESULTS

In a population of 100,000 people, moving from current to best practice for beta-adrenoceptor antagonists and smoking cessation will prevent 11 and 4 deaths (or gain of 127 or 42 life-years), respectively. The cost per event prevented in the next year, or life-year gained, is less for beta-adrenoceptor antagonists than for smoking cessation. Public health professionals were found to be more inclined to rank alternative interventions according to the population cost impact than the QALY approach.

DISCUSSION

The use of the population cost-impact approach allows information on the benefits of moving from current to best practice to be presented in terms of the benefits and costs to a particular population. The process for deciding between alternative interventions in a prioritisation exercise may differ according to the local context. We suggest that the valuation of the benefit is performed after the benefits have been quantified and that it takes into account local issues relevant to prioritisation. It would be an appropriate next step to experiment with, and formalise, this part of the population cost-impact analysis to provide a standardised approach for determining willingness to pay and provide a ranking of priorities.

CONCLUSION

Our method adds a new dimension to economic analysis, the ability to identify costs and benefits of potential interventions to a defined population, which may be of considerable use for policy makers working at the local level.

Hormone therapy and radiotherapy for early prostate cancer: a utility-adjusted number needed to treat (NNT) analysis

PURPOSE

To quantify, using the number needed to treat (NNT) methodology, the benefit of short-term (< or =6 months) hormone therapy adjuvant to radiotherapy in the group of patients with early (clinical stage T1-T2c) prostate cancer.

METHODS AND MATERIALS

The absolute biochemical control benefit for the use of hormones adjuvant to radiotherapy in early-stage disease was determined by literature review. A model was developed to estimate the utility-adjusted survival detriment due to the side effects of hormone therapy. The NNTs before and after the incorporation of hormone sequelae were computed; the sign and magnitude of the NNTs were used to gauge the effect of the hormones.

RESULTS

The absolute NNT analysis, based on summarizing the results of 8 reports including a total of 3652 patients, demonstrated an advantage to the addition of hormones for the general early-stage prostate cancer population as well as for all prognostic groups. After adjustment for hormone-induced functional loss, the advantage of hormones remained considerable in the high- and intermediate-risk groups, with the utility-adjusted NNT becoming weakened in the low-risk group when the utility compromise from complications of hormones was assumed to be considerable.

CONCLUSIONS

Short-term hormone therapy seems to be beneficial for selected early-stage prostate cancer patients. The advantage seems to be greatest in the intermediate- and high-risk groups; with current follow-up, the side effects of hormones may outweigh their benefit in certain clinical situations in the favorable group. The present investigation demonstrates the significant role of the NNT technique for oncologic and radiotherapeutic management decisions when treatment complications need to be considered and balanced with the beneficial effects of the treatment.

The number needed to treat (NNT) can be adjusted for bias when the outcome is measured with error

BACKGROUND AND OBJECTIVE

We consider the number needed to treat (NNT) when the event of interest is defined by dichotomizing a continuous response at a threshold level. If the response is measured with error, the resulting NNT is biased. We consider methods to reduce this bias.

METHODS

Bias adjustment was studied using simulations in which we varied the distributions of the underlying response and measurement error, including both normal and nonnormal distributions. We studied a maximum likelihood estimate (MLE) based on normality assumptions, and also considered a simulation-extrapolation estimate (SIMEX) without such assumptions. The treatment effect across all potential thresholds was summarized using an NNT threshold curve.

RESULTS

Crude NNT estimation was substantially biased due to measurement error. The MLE performed well under normality, and it continued to perform well with nonnormal measurement error, but when the underlying response was nonnormal the MLE was unacceptably biased and was outperformed by the SIMEX estimate. The simulation results were also reflected in empirical data from a randomized study of cholesterol-lowering therapy.

CONCLUSION

Ignoring measurement error can lead to substantial bias in NNT, which can have an important practical effect on the interpretation of analyses. Analysis methods that adjust for measurement error bias can be used to assess the sensitivity of NNT estimates to this effect.

A Method to Enhance Determining the Clinical Relevance of Periodontal Research Data: Number Needed to Treat (NNT)

It would be advantageous if clinical trials reported both statistical and clinically meaningful results. In this regard, determination of the number of sites that would need to be treated in a test group to provide a beneficial result or prevent an adverse event at one additional site beyond the control group would provide useful information. This editorial addresses the use of NNT (number needed to treat) calculations to enhance determining the clinical relevance of periodontal research findings. The application, requirements for use, benefits, and limitations of employing NNT calculations are discussed.

Absolute benefit, number needed to treat and gain in life expectancy: which efficacy indices for measuring the treatment benefit?

The absolute benefit (AB) is extensively used to summarize the results of clinical trials. As the AB depends directly on the patient's baseline risk, therapeutic decisions based on AB tend to favor patients at high risk. To evaluate the consequences of this decision's procedure for life-long therapy, we compare the AB with the gain in event-free life expectancy in a simulated hypertensive population. Our results show that the AB goes through a maximum and then declines as the duration of treatment increases. The amplitude of the variation of AB is independent of the baseline risks but the maximum is reached more quickly in the high-risk patients. Considering the gain in event-free life expectancy, low-risk patients benefit more than high-risk patients do, at the expense of a longer treatment exposure. The interpretation of the AB changes depending on follow-up.

Hormone therapy adjuvant to external beam radiotherapy for locally advanced prostate carcinoma

BACKGROUND

Hormone therapy commonly is used to treat metastatic, locally advanced, and localized prostate carcinoma. The objective of the current investigation was to determine, using the number-needed-to-treat (NNT) method, the effect of using hormone therapy to treat locally advanced disease, with consideration given to both the complications and the known advantages associated with hormone therapy.

METHODS

A literature review was performed to determine 1) the absolute benefit, based on available clinical endpoints, associated with the addition of hormone therapy to external beam radiotherapy for locally advanced prostate carcinoma; 2) the incidence of side effects of short-term and long-term hormone therapy; and 3) the stepwise progression from biochemical failure to death. A model was constructed to estimate the complication/utility-adjusted survival detriment resulting from the side effects of short-term (</= 6 months) and long-term (> 6 months) hormone therapy, and the absolute/unadjusted and complication-adjusted NNTs for the addition of short-term and long-term hormone therapy were computed. In all cases, the magnitudes and signs of the NNTs obtained were used to gauge the effect of hormone therapy.

RESULTS

The unadjusted NNTs were positive and in most cases had relatively small magnitudes (the greater the NNT, the smaller the benefit) for both short-term and long-term hormone therapy; these results were expected, and they suggested that there is a strong benefit associated with the use of hormones adjuvant to radiotherapy for locally advanced disease. Adjusted NNTs remained positive and had relatively small magnitudes even after the introduction into the analysis of complications of short-term and long-term hormone therapy. This finding, although weak with respect to the effect of short-term hormone therapy on cause-specific survival, remained robust over the range of values for utility impairment expected from short-term and long-term hormone therapy.

CONCLUSIONS

The benefits of short-term and long-term hormone therapy for locally advanced prostate carcinoma appear to be significant and to outweigh the associated side effects. Long-term therapy appears to be better than short-term therapy in terms of virtually all endpoints studied, even when the increased incidence of side effects is considered. The current investigation was successful in the use of the complication-adjusted NNT method for oncologic and radiotherapeutic scenarios in which the results of randomized trials could be summarized, adjusted for treatment toxicity, and individualized to a given patient.

Calculating the "number needed to be exposed" with adjustment for confounding variables in epidemiological studies

The number needed to treat (NNT) is a popular summary statistic to describe the absolute effect of a new treatment compared with a standard treatment or control concerning the risk of an adverse event. The NNT concept can be applied whenever the risk of an adverse event is compared between two groups; for the comparison of exposed with unexposed subjects in epidemiological studies, we propose the term "number needed to be exposed" (NNE). Whereas in randomized clinical trials NNT can be calculated on the basis of a simple 2 x 2 table, in epidemiological studies methods to adjust for confounders are required in most applications. We derive a method based upon multiple logistic regression analysis to perform point and interval estimation of NNE with adjustment for confounding variables. The adjusted NNE can be calculated from the adjusted odds ratio (OR) and the unexposed event rate (UER) estimated by means of an appropriate multiple logistic regression model. As UER is dependent on the confounders, the adjusted NNEs also vary with the values of the confounding variables. Two methods are proposed to take the dependence of NNE on the values of the confounders into account. The adjusted number needed to be exposed can be a useful complement to the commonly presented results in epidemiological studies, such as ORs and attributable risks.