Microsoft word - 210610-statistics review.doc

Detail-Document #210610
−This Detail-Document accompanies the related article published in− PHARMACIST'S LETTER / PRESCRIBER'S LETTER
June 2005 Volume 21 Number 210610 Applying Study Results to Patient Care:
Relative Risk, Absolute Risk, and Number Needed to Treat
Lead author: Jill Allen, Pharm.D., BCPS The manner in which study results are Clinical trials evaluating the safety and presented affects the way they are viewed. efficacy of drug therapy often use three related Clinicians are more interested in results that are statistical methods to report results: relative risk, portrayed as large whole numbers. A recent study relative risk reduction, and odds ratio. These illustrates this point.1 Clinicians were presented terms can also be used to calculate two very with study results in four different formats: practical clinical tools: the number needed to treat Format A: 91.8% survival with active (NNT) and the number needed to harm (NNH). treatment vs. 88.5% survival with placebo. Format D above illustrates the NNT. Format B Format B: Active treatment led to a 30% illustrates relative risk reduction. Format C reduction in mortality. illustrates absolute risk reduction. Portraying Format C: Active treatment reduced mortality results as relative rather than absolute risk reduction can make a drug's efficacy appear more Format D: One death was avoided for every impressive. This is why pharmaceutical 30 patients treated. marketing often focuses on relative risk.2 While 70% of clinicians would implement the results of Formats B and D in their practice, only statistical tools helps clinicians make more 20% would act on the Formats A and C. In informed choices about drug therapy and makes reality, all four formats present the results of the them less susceptible to pharmaceutical marketing same study, the milestone 4S study demonstrating methods. (See box on next page for calculations cardiovascular risk reduction with simvastatin.1 of above examples). Simvastatin
Absolute risk reduction 11.5% – 8.2%=3.3%*
Risk in treatment group divided by risk in control group 8.2% ÷ 11.5%=0.71 Relative risk reduction Absolute risk reduction divided by risk in control group
3.3 ÷ 11.5=0.29 or 29%* OR
1 minus relative risk (1 – 0.71=0.29 or 29%)
Number needed to treat 1 divided by absolute risk reduction 1 ÷ 3.3=30 patients treated to avoid one death * Actual calculated numbers differ slightly from examples presented by O'Connell et al.1
Relative Risk
determine ahead of time whether patients will Relative risk compares the risk of an event in receive active treatment or control.4,5 individuals with a particular characteristic to the A relative risk of 1 indicates no association risk of that event in individuals without that between treatment and outcome. A relative risk characteristic. In a clinical trial, this would be the greater than 1 indicates a positive association outcome in the treatment group divided by the between treatment and outcome. A relative risk outcome in the control group.3 Relative risk can less than 1 indicates a negative association only be used in prospective cohort studies between treatment and outcome.6,7 A study because, by definition, it requires that you investigating an anticoagulant for prevention of thrombosis might use relative risk to portray both Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 6 of 14)
efficacy and safety. For efficacy, a relative risk in case-control studies. Case-control studies less than 1 might indicate a decreased risk of compare patients with an outcome of interest to thrombosis. In terms of side effects, a relative patients without that outcome. This type of study risk greater than 1 might indicate an association is often used to determine whether drugs are the between the anticoagulant and bleeding. cause of rare adverse events. The odds of Relative and Absolute Risk Reduction
exposure to the suspected drug is compared in cases who have the adverse event and controls Relative risk reduction is 1 minus the relative who do not have the adverse event. Odds ratios risk.3 Portraying the benefits of treatment as and relative risk provide comparable estimates of relative risk reduction can mislead clinicians risk when the outcome is rare. But the odds ratio about the value of that treatment unless they can exaggerate risk when the disease or outcome consider the patient's baseline risk for the is common (incidence greater than 10%).4,5,7 The outcome the treatment is preventing. For odds ratio cannot be used directly to calculate an example, when deciding whether to prescribe a NNT, but it can be done using standard formulas drug to prevent myocardial infarction, one should and nomograms.9 One such nomogram can be consider the patient's baseline risk of myocardial viewed at http://www.cebm.net/nnts.asp. An interactive tutorial prepared by Chris Cates, Number Needed to Treat and Harm
a general practitioner with a talent for The NNT and NNH are statistical concepts that demystifying evidence-based medicine, illustrates share the simplicity of relative risk reduction, but this concept very clearly.8 He considers the they have less potential to be misleading because decision of whether to prescribe clopidogrel in they are based on absolute risk. These very addition to aspirin based on results of the CURE understandable terms can help both clinicians and trial. The relative risk reduction for vascular patients decide whether the risks and benefits of events with clopidogrel is 20%. The absolute risk treatment are worthwhile. The NNT is the reduction in the CURE trial is 2.1% -- from 11.4% reciprocal of the absolute risk reduction with to 9.3%. An individual patient's risk for vascular drug treatment (1 divided by absolute risk events might vary from that of patients in the reduction).3,7 In clinical trials of drug therapy, CURE trial. If the patient's baseline risk of a it is the number of patients who would need vascular event is 15%, treatment with clopidogrel to be treated in order to achieve benefit in will reduce that patient's absolute risk of a one patient. The NNH is the reciprocal of vascular event to 12%. If the patient's baseline the absolute risk increase with a drug side risk of a vascular event is only 1%, treatment with effect. In other words, it is the number of clopidogrel will only reduce that patient's patients who would be treated before you absolute risk of an event to 0.8%.8 expect to see one patient with an adverse Just as relative risk can make treatment look effect. Comparing the NNT and NNH can more effective, it can make adverse effects appear help give an accurate assessment of the risks more frightening. Stephen Gehlbach illustrates and benefits of treatment. this point with the following example. In the Dr. Cates illustrates this point with the results 1970's, oral contraceptives were found to increase of a Cochrane review evaluating antibiotics for the risk of myocardial infarction by 2.5- to 5-fold. the treatment of pediatric otitis media. The This statistic sounds very alarming until one primary benefit of treatment is pain relief two to considers that this is an absolute risk of 3.5 deaths seven days after antibiotics are begun. Pain per 100,000 users per year.4 resolves quickly in most children even without antibiotic therapy. Pain tends to persist longer in Odds Ratio
younger children. In general, 15 children need to The odds of an event is the ratio of the number be treated with antibiotics to relieve pain in one of events to the number of non-events (similar to child (NNT=15). For children under two years of the way the odds of winning or losing a horse race age, the NNT is nine. The primary risk of is expressed at a race track).5 The odds ratio is antibiotic therapy is side effects. Only 12 children the odds of exposure in cases divided by the odds need to be treated for one child to develop of exposure in controls.9 It is analogous to vomiting, rash, or diarrhea (NNH=12).10 relative risk.7 Unlike relative risk, it can be used Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 7 of 14)
Dr. Cates has developed a user-friendly the case with moderate hypertension, treatment computer program that will calculate the NNT provides a relative risk ratio of 0.6 (0.009/0.015) from a meta-analysis of drug therapy, particularly and a relative risk reduction of 40% (1–0.6=0.40). Cochrane reviews. It is called Visual Rx and can The absolute risk reduction is much lower (0.015– be accessed from http://www.nntonline.net/. 0.009=0.006). The number needed to treat in this case is 167 (1/0.006=166.66). In other words, 167 patients would need to be treated for five years to prevent one stroke.3 There is always some uncertainty about how How to Go from Absolute Risk to NNT
well the NNT represents the true treatment effect Cook and Sackett use the treatment of mild to in the population at large. This uncertainty can be moderate hypertension to illustrate the relation- expressed as a confidence interval. A confidence ship between relative risk, absolute risk, and the interval estimates the range within which the true number needed to treat.3 About 20% of patients treatment effect lies. A narrow confidence interval suggests less uncertainty and a wide expected to have a stroke over a five-year period. confidence interval suggests more uncertainty. Antihypertensive therapy reduces this risk to 12%. Ideally, a NNT for drug therapy should be This provides a relative risk ratio of 0.6 (0.12/0.2) accompanied by information about what it was and a relative risk reduction of 40% (1– compared to (another drug or placebo), the 0.60=0.40). This is an absolute risk reduction of duration of treatment, the study outcome, and a 8% (0.20–0.12=0.08). The reciprocal of absolute 95% confidence interval.12 We have compiled risk (1/0.08) is the number needed to treat, in this NNTs for drug therapy of common disorders in case approximately 13. Thirteen patients would the table below. Some of these NNTs are based need to be treated with antihypertensive therapy on a single large-scale clinical trial, while others for five years to prevent one stroke.3 are based on a systematic review or meta-analysis They take this example a step further and of multiple clinical trials. When comparing the compare how treatment reduces the risk of stroke NNT of two drug regimens, make sure that they in patients with mild hypertension. Over a 5-year are based on the same duration of therapy, treat period, 1.5% of patients with untreated mild the same condition, and share the same outcome.14 hypertension would have a stroke compared with 0.9% of antihypertensive-treated patients. As is Drug Therapy of Common Conditions and the Number Needed to Treat*
Condition
Number Needed to Treat or Harm
Duration of therapy
(95% confidence interval)
Coronary artery disease
Primary prevention of CHD Aspirin x 1 year
500 healthy men treated to prevent one MI/death28 Statin x 3 to 5 years 71 treated to prevent one MI/stroke11 Coronary artery disease28 ACE inhibitor x 1 year 22 to 83 treated to prevent one death Beta blocker x 1 year 31 to 81 treated to prevent one death Simvastatin x 1 year 163 treated to prevent one death Unstable angina28 Aspirin x 1 year 25 to prevent one MI/death Myocardial infarction28 Streptokinase + 1 NNT: 20 treated to prevent one death at 5 weeks NNH: 1000 treated to cause one hemorrhagic stroke tPA vs. streptokinase 100 treated to prevent one extra death 18 treated to prevent 1 death within 6 months Intensive lipid-lowering Target of 70 mg/dL 50 extra patients treated per year to 70 mg/dL rather than after acute coronary (atorvastatin) vs. 100 100 mg/dL to prevent one CHD event mg/dL (pravastatin) Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 8 of 14)
Secondary prevention of Simvastatin x 5 years 15 (10-25) to prevent one major coronary event 29 (18-56) to prevent one coronary death14 Statin x 5 years 21 treated to prevent 1 MI/stroke11 Prevention of CHD events Statin x 15 years Number of patients treated to prevent 1 CHD event in elderly patients with 10-year risk of 10%: 10 hyperlipidemia, based on 10-year risk of 20%: 5 10-year risk of MI or 10-year risk of 30%: 3 coronary death22 10-year risk of 40%: 2 Hypertension (HTN)
Mild HTN14
Antihypertensive x 1 700 treated to prevent one stroke, MI, or death Mild HTN (10-year CHD Antihypertensive x 5 40 treated to prevent one cardiovascular complication risk of at least 15%)23 years Aspirin x 5 years 90 treated to prevent 1 cardiovascular complication Severe hypertension14 Antihypertensive x 1 15 treated to prevent 1 stroke, MI, or death HTN in elderly28 Antihypertensive x 5 18 treated to prevent 1 cardiovascular complication Isolated systolic HTN28 43 treated to prevent 1 stroke atenolol x 1 year HTN in diabetes28 Antihypertensive x 10 15 treated to prevent 1 diabetes-related death Heart failure
Heart failure, NYHA I-II14
ACE inhibitor x 1 year 100 treated to prevent 1 death Heart failure, NYHA IV14 ACE inhibitor x 1 year 6 treated to prevent 1 death Heart failure post-MI14 18 treated to prevent 1 death Heart failure, NYHA II-IV32 Metoprolol ER 25 treated to prevent 1 death 50 treated to prevent 1 death Thromboembolic events
Deep vein thrombosis31
Low molecular weight NNT: 61 to avoid 1 death; 114 to avoid 1 recurrent heparin vs. heparin thromboembolism with heparin. NNT: 164 to avoid 1 major bleed with heparin Condition
Number Needed to Treat or Harm
Duration of therapy
(95% confidence interval)
Stroke
Prevention of stroke in
Warfarin, primary 37 treated to prevent 1 major vascular event atrial fibrillation29 prevention x 1 year Warfarin, secondary 13 treated to prevent 1 major vascular event prevention x 1 year Aspirin, primary 67 treated to prevent 1 major vascular event prevention x 1 year Aspirin, secondary 40 treated to prevent 1 major vascular event prevention x 1 year Primary prevention of Pravastatin x 1 year 641 patients with hyperlipidemia treated to prevent 1 Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 9 of 14)
Secondary prevention of Smoking cessation x 1 43 to prevent 1 major vascular event year Aspirin x 1 year 38 to prevent 1 stroke after TIA or minor stroke28 100 to prevent 1 major vascular event Antihypertensive x 1 42 to 45 treated to prevent 1 major vascular event year Statin x 1 year 59 treated to prevent 1 major vascular event Acute ischemic stroke29 Thrombolytic (tPA) 7 treated to improve outcome in 1 patient Modification of Cardiovascular Risk Factors
Smoking cessation14
Nicotine gum, patch, 14 treated for 1 success over 6 to 12 months of follow-up spray, or inhaler Weight reduction in Sibutramine x 6 months 2.7 treated for 1 to have 5% weight reduction Orlistat x 1 year 3.9 treated for 1 to have 5% weight reduction Orlistat x 1 year 5.6 treated for 1 to have 10% weight reduction Dermatologic conditions
Athletes foot17
2 treated to achieve one extra cure Undecylenic acid or 2 treated to achieve one extra cure Self-administered 4 (3-12) treated for 1 cure Terbinafine 250 mg vs. 2.7 treated with terbinafine for 1 extra patient with cured griseofulvin 500 mg x 12 weeks Terbinafine x 16 weeks 2.5 treated with terbinafine for 1 extra patient with cured vs. griseofulvin 500 mg x 52 weeks Terbinafine x 24 weeks 4.6 treated with terbinafine for 1 extra patient with cured vs. griseofulvin 1000 mg x 48 weeks Condition
Number Needed to Treat or Harm
Duration of therapy
(95% confidence interval)
Endocrine Disorders
Prevention of type 2
7 treated to prevent 1 case in 3 years 14 treated to prevent 1 case in 3 years Treatment of type 2 Metformin x 1 year Obese patients: 141 treated to prevent 1 death; 74 treated to prevent 1 diabetes-related outcome Tight blood pressure 152 to prevent 1 diabetes-related death; control x 1 year 61 treated to prevent 1 complication Tight glucose control x 196 patients treated to prevent 1 complication 1 year Aspirin, primary 45 treated to prevent 1 major cardiovascular event28 prevention x 5 years Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 10 of 14)
Simvastatin x 5 years 6 patients with known CHD treated to prevent 1 major cardiovascular event28 Polycystic ovary disease25 4 women treated for 1 to achieve ovulation Postmenopausal hormone Premarin plus NNT for 5 years: 333 to prevent 1 hip fracture; replacement therapy35 medroxyprogesterone 333 to prevent 1 colorectal cancer NNH for 5 years of treatment: 250 to cause 1 CHD event; 250 to cause 1 stroke; 100 to cause 1 venous thromboembolism; 200 to cause 1 breast cancer Gastrointestinal disorders
Prevention of GI
Misoprostol x 1 year 83 treated to prevent 1 serious GI complication; NNT as complications with NSAIDs low as 7 for age over 75 years + history of GI bleed Misoprostol 800 mcg x 6 treated to prevent one GI complication 6 months Omeprazole 20 mg x 6 3 treated to prevent one GI complication Prevention of GI events Rofecoxib vs. naproxen 41 treated with rofecoxib instead of naproxen to avoid 1 with coxib over traditional upper GI complication† Celecoxib vs. NSAID x 100 treated with celecoxib instead of NSAID to avoid 1 upper GI complication† GERD, symptom relief 21 For excellent/good symptom relief in 1 patient: 14 patients treated with either; 6 treated with both GERD, short-term healing For every 3 treated with omeprazole, 1 extra patient ranitidine x 8 weeks healed than would have healed with ranitidine For every 3 treated with omeprazole, 1 extra patient still ranitidine x 1 year healed at 1 year than expected with ranitidine Postoperative nausea and 7 treated to prevent nausea in 1 5-6 treated to prevent nausea in 1 Peptic ulcer disease14 Triple antibiotics vs. NNT for H. pylori eradication is 1.1 at 6 weeks and 1.8 at 1 year; NNT is 5 for ulcer healing at 6 weeks Condition
Number Needed to Treat or Harm
Duration of therapy
(95% confidence interval)
Infectious Diseases
Pediatric ear infections10
NNT: 15 treated to relieve pain in 1 NNT for <2 year old: 9 treated to relieve pain in 1 NNH: 12 treated to cause 1 case of vomiting, rash, or diarrhea 23 immunized to prevent 1 case of influenza Prophylaxis of infection 16 (9-92) treated to prevent 1 infection after dog bite12 Streptococcal pharyngitis26 3000-4000 patients treated to prevent 1 case of acute rheumatic fever Ipratropium nasal For 1 patient to have improvement in runny nose, the NNT is 6.3 vs saline and 1.6 vs no treatment 3 treated for 1 to have cold symptoms resolved between days 6 to 12 Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 11 of 14)
Neurology
Dementia14
8 treated for 1 to have 4-point improvement on ADAS-cog Multiple sclerosis, Interferon beta-1b x 2 9 treated to prevent confirmed progression in 1; secondary progressive14 11 treated to prevent 1 moderate/severe relapse; 13 treated to prevent 1 becoming wheel-chair bound Multiple sclerosis, Interferon beta-1a x 2 5 patients treated to prevent 1 moderate/severe relapse remitting-relapsing14 Pain
Acute migraine14
PO sumatriptan 100 mg 3 treated for one 2-hour headache response SC sumatriptan 6 mg 2 treated for one 2-hour headache response PO eletriptan 80 mg 2.6 treated for one 2-hour headache response 3.7 treated for one to be pain-free at 2 hours 2.8 for 1 response sustained at 24 hours PO eletriptan 40 mg 2.9 treated for one 2-hour headache response 4.5 treated for 1 to be pain-free at 2 hours 3.6 for 1 response sustained at 24 hours PO eletriptan 20 mg 4.4 treated for one 2-hour headache response 9.9 treated for 1 to be pain-free at 2 hours 5.4 for 1 response sustained at 24 hours PO rizatriptan 10 mg 2.7 treated for one 2-hour headache response 3.1 treated for 1 to be pain-free at 2 hours 5.6 for 1 response sustained at 24 hours PO rizatriptan 5 mg 3.9 treated for one 2-hour headache response 4.7 treated for 1 to be pain-free at 2 hours 8.3 for 1 response sustained at 24 hours 3.9 treated for one 2-hour headache response Neuropathic pain For 1 patient with at least 50% reduction in pain: Treat 3 with diabetic neuropathy12, 27 Treat 4 (2.6-8.9) with postherpetic neuralgia15 Topical capsaicin 3-6 treated for 1 to experience pain relief14,27 3 (1.9-4.2) treated for 1 with at least 50% pain relief15 3 (2.4-8.7) treated for 1 to experience pain relief 27 Condition
Number Needed to Treat or Harm
Duration of therapy
(95% confidence interval)
Pain (cont.)
Acute pain14
Celecoxib 200 mg 2.8 (2.1 to 4.4) for 1 with at least 50% pain reduction 1.9 (1.6 to 2.2 ) for 1 with at least 50% pain reduction Ibuprofen 400 mg 2.1 (1.7 to 2.6) for 1 with at least 50% pain reduction Postoperative pain, 2-3 treated for 1 with at least 50% pain reduction moderate to severe14 1.7 treated for 1 with at least 50% pain reduction IM morphine 10 mg 2.9 treated for 1 with at least 50% pain reduction PO APAP 650 mg + 3 treated for 1 with at least 50% pain reduction codeine 60 mg IM ketorolac 30 mg 3.4 treated for 1 with at least 50% pain reduction IM ketorolac 10 mg 5.7 treated for 1 with at least 50% pain reduction 4.5 treated for 1 with at least 50% pain reduction PO tramadol 75 mg 5 treated for 1 with at least 50% pain reduction Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 12 of 14)
Osteoarthritis14 5 treated for improved symptoms in 1 Topical capsaicin 3 treated for pain relief in 1 Prevention of hip fracture in Calcium1200 mg + 14 treated to prevent any fracture; 20 to 40 treated to ambulatory elderly14 vitamin D x 3 years prevent 1 hip fracture Rheumatoid arthritis30 2 treated for 1 extra patient to achieve ACR20; 4 treated for 1 extra patient to achieve ACR50; 8 treated for 1 extra patient to achieve ACR70. 4 treated for 1 patient to achieve ACR20 5-6 treated for 1 extra patient to achieve ACR50 4 treated for 1 patient to achieve ACR20 5-6 treated for 1 extra patient to achieve ACR50 Severe postmenopausal Bisphosphonate x 3 9 women treated to prevent 1 new spinal fracture years (risedronate) Urology
Benign prostatic
Finasteride x 2 years 26 to 38 men treated to prevent prostatectomy or acute urinary retention16 Finasteride + alpha 9 men treated to prevent clinical progression in 1 (based blocker x 4 years Erectile dysfunction, mixed 2 men treated for 1 to have erection suitable for etiology/diabetes14 Abbreviations: ACR20 = 20% reduction in American College of Rheumatology criteria; ACR50 = 50% reduction
in American College of Rheumatology criteria; ACR70 = 70% reduction in American College of Rheumatology
criteria; CVD = coronary vascular disease; HTN = hypertension; LVD = left ventricular dysfunction; MI =
myocardial infarction; NYHA = New York Heart Association.
* Unless stated in the table, NNTs are based on comparisons of drug regimens with placebo.
† Differences in NNT between rofecoxib and celecoxib may be due to differences in the population in which they
were studied.

Users of this document are cautioned to use their own _00108.asp. (Accessed March 20, 2004). professional judgment and consult any other necessary 6. Bigby M. Odds ratios and relative risks. Arch or appropriate sources prior to making clinical Dermatol 2000;136:770-1. judgments based on the content of this document. Our 7. Bjornson DC. Interpretation of drug risk and benefit: Individual and population perspectives. editors have researched the information with input Ann Pharmacother 2004;38:694-9. from experts, government agencies, and national 8. Cates C. Understanding statistics: an interactive organizations. Information and Internet links in this article were current as of the date of publication. http://www.bmjlearning.com. (Accessed March 20, 9. McAlister FA, Straus SE, Guyatt GH, et al. Users' guides to the medical literature: XX. Integrating research evidence with the care of the individual patient. JAMA 2000;283:2829-36. 1. O'Connell RL, Gebski VJ, Keech AC. Making 10. Cates C. Dr. Chris Cates' EBM Web Site. sense of trial results: outcomes and estimations. Med J Aust 2004;180:128-30. (Accessed March 22, 2004). 2. Lexchin J. How patient outcomes are reported in 11. Therapeutics initiative: Evidence Based Drug drug advertisements. Can Fam Physician 1999;45: Therapy. Do statins have a role in primary 3. Cook RJ, Sackett DL. The number needed to treat: www.ti.ubc.ca/PDF/48.pdf. (Accessed March 22, a clinically useful measure of treatment effect. BMJ 12. Moore A, McQuay HJ. What is an NNT? Available 4. Gehlback SH. Interpreting the medical literature. 3rd ed. 1993 McGraw-Hill. NY. medicine.co.uk. (Accessed March 22, 2004). 5. Last A, Wilson S. Relative risk and odds ratio: 13. Cannon CP, Braunwald E, McCabe C, et al. What's the difference? J Fam Pract 2004;53(2). Intensive versus moderate lipid lowering with Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 13 of 14)
statins after acute coronary syndromes. N Engl J 29. European Stroke Initiative Executive Committee Med 2004;350:1495-504. Available online at: and the EUSI Writing Committee. European Stroke http://content.nejm.org/. (Accessed March 20, management - Update 2003. Cerebrovasc Dis 14. Bandolier: "Evidence based thinking about health 30. Jobanputra P, Barton P, Bryan S, Burls A. The effectiveness of infliximab and etanercept for the (Accessed March 20, 2004). treatment of rheumatoid arthritis: a systematic 15. Raja SN, Haythornthwaite JA, Pappagallo M, et al. review and economic evaluation. Health Technol Opioids versus antidepressants in postherpetic neuralgia: a randomized, placebo-controlled trial. http://www.hta.nhsweb.nhs.uk. (Accessed March Neurology 2002;59:1015-21. 16. Barry MJ, Roehrborn CG. Benign prostatic 31. Hirsh J, Anand SS, Halperin JL, Fuster V. Guide to hyperplasia. BMJ 2001:323:1042-6. anticoagulant therapy: Heparin: a statement for 17. Crawford F, Hart R, Bell-Syer SE, et al. Extracts healthcare professionals from the American Heart from "Clinical evidence": Athlete's foot and fungally Association. Circulation 2001;103:2994-3018. infected toenails. BMJ 2001:322:288-9. 32. Tangeman HJ, Patterson JH. Extended-release 18. Bigby M. At what rates do commonly used local metoprolol succinate in chronic heart failure. Ann treatments lead to complete disappearance of the Pharmacother 2003;37:701-10. treated wart? Arch Dermatol 2003;139:801-2. 19. American Diabetes Association, National Institute of Diabetes, Digestive, and Kidney Diseases. The prevention or delay of type 2 diabetes. Diabetes Care 2003;26 (suppl 1):S62-9. 20. Shaughnessy AF, Slawson DC. What happened to the valid POEMs? A survey of review articles on the http:www.bmj.org. 21. Peterson, WL and American Gastroenterological Improving the management of GERD. Evidence-based therapeutic strategies. 2002. Available online http://www.gastro.org/edu/GERDmonograph.pdf. (Accessed March 27, 2004). 22. National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Adult Treatment Panel III Report. 2002. Available online at: http://www.nhlbi.nih.gov/guidelines/cholesterol/index.htm. (Accessed March 27, 2004). 23. Ramsay L, Williams B, Johnston G, et al. Guidelines for management of hypertension: report of the third working party of the British Hypertension Hypertens postoperative nausea and vomiting. Expert Opin Pharmacother 2003;4:457-73. 25. Lord JM, Flight IH, Norman RJ. Metformin in polycystic ovary syndrome: systematic review and meta-analysis. BMJ 2003;327:951-3. 26. Cooper RJ, Hoffman JR, Bartlett JG et al. Principles of appropriate antibiotic use for acute pharyngitis in adults: background. Ann Intern Med 2001;134:509-17. 27. Koltzenburg M, Scadding, J. Neuropathic pain. Curr Opin Neurol 2001;14:641-7. 28. InfoPOEMS InfoRetriever. Available online at: http://www.infopoems.com. (Accessed March 27, 2004). Copyright 2005 by Therapeutic Research Center Pharmacist's Letter / Prescriber's Letter P.O. Box 8190, Stockton, CA 95208 Phone: 209-472-2240 Fax: 209-472-2249 (Detail-Document #210610: Page 14 of 14)
33. Bjorkman DJ. Commentary: Gastrointestinal safety 35. World Health Organization. WHO Drug Information of coxibs and outcome studies: What's the verdict? J Pain Symptom Manage 2002;23 (suppl 4):S11-4. 34. Pitt B, Remme W, Zannad F, et al. Eplerenone, a 3/vol17-1.pdf. (Accessed April 25, 2004). selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N Engl J Med 2003;348:1309-21. The most practical knowledge in the least time… 3120 West March Lane, P.O. Box 8190, Stockton, CA 95208 TEL (209) 472-2240 FAX (209) 472-2249 Copyright  2005 by Therapeutic Research Center Subscribers to Pharmacist's Letter and Prescriber's Letter can get Detail-Documents, like this one, on any topic covered in any issue by going to www.pharmacistsletter.com or www.prescribersletter.com

Source: http://www.ulm.edu/~biglane/APPE%20Relative%20Risk,%20Absolute%20Risk,%20and%20Number%20Needed%20to%20Treat.pdf