Biostat.jhsph.edu

Drug-Drug Interactions Among Elderly
Patients Hospitalized for Drug Toxicity
David N. Juurlink, MD, FRCPC
Context Drug-drug interactions are a preventable cause of morbidity and mortality,
yet their consequences in the community are not well characterized.
Objective To determine whether elderly patients admitted to hospital with specific
drug toxicities were likely to have been prescribed an interacting drug in the week
Design Three population-based, nested case-control studies.
Setting Ontario, Canada, from January 1, 1994, to December 31, 2000.
DVERSE DRUG EVENTS AFFECTmillions of patients each year Patients All Ontario residents aged 66 years or older treated with glyburide, di-
goxin, or an angiotensin-converting enzyme (ACE) inhibitor. Case patients were those
admitted to hospital for drug-related toxicity. Prescription records of cases were com- pared with those of controls (matched on age, sex, use of the same medication, and presence or absence of renal disease) for receipt of interacting medications (co- trimoxazole with glyburide, clarithromycin with digoxin, and potassium-sparing di- Main Outcome Measure Odds ratio for association between hospital admission
for drug toxicity (hypoglycemia, digoxin toxicity, or hyperkalemia, respectively) and use of an interacting medication in the preceding week, adjusted for diagnoses, re- ceipt of other medications, the number of prescription drugs, and the number of hos- pital admissions in the year preceding the index date.
larly important type of adverse drug event Results During the 7-year study period, 909 elderly patients receiving glyburide
were admitted with a diagnosis of hypoglycemia. In the primary analysis, those on previous reports, clinical studies, and patients admitted for hypoglycemia were more than 6 times as likely to have been treated with co-trimoxazole in the previous week (adjusted odds ratio, 6.6; 95% confidence interval, 4.5-9.7). Patients admitted with digoxin toxicity (n = 1051)were about 12 times more likely to have been treated with clarithromycin (adjusted odds ratio, 11.7; 95% confidence interval, 7.5-18.2) in the previous week, and patients treated with ACE inhibitors admitted with a diagnosis of hyperkalemia tions from the marketplace.4-10 Such dras- (n = 523) were about 20 times more likely to have been treated with a potassium- tic measures are probably justifiable be- sparing diuretic (adjusted odds ratio, 20.3; 95% confidence interval, 13.4-30.7) in the previous week. No increased risk of drug toxicity was found for drugs with simi- serious drug-drug interactions.11-15 More- lar indications but no known interactions (amoxicillin, cefuroxime, and indapamide, Conclusions Many hospital admissions of elderly patients for drug toxicity occur af-
tions fail to detect up to a third of drug- ter administration of a drug known to cause drug-drug interactions. Many of these drug interactions, while frequently alert- interactions could have been avoided.
ing pharmacists to trivial or nonspecific Author Affiliations: Sunnybrook and Women’s Col-
improvement services, and the University of Arizona lege Health Sciences Centre; the Clinical Epidemiol- on a project funded by the Centers for Disease Con- tions in clinical practice, and most of the ogy and Healthcare Research Program, and Depart- trol and Prevention examining the clinical signifi- ments of Medicine (Drs Juurlink, Laupacis, and cance of, and management strategies for, various drug- Redelmeier), and Pharmacy (Dr Mamdani), Univer- sity of Toronto; and the Institute for Clinical Evalua- Corresponding Author and Reprints: David Juurlink,
tive Sciences (Drs Juurlink, Mamdani, Laupacis, and MD, FRCPC, G Wing 106, Sunnybrook and Women’s pitalized patients.18-23 No studies to date Redelmeier, and Mr Kopp), Toronto, Ontario.
College Health Sciences Centre, 2075 Bayview Ave, Financial Disclosure: Dr Juurlink has served as a
Toronto, Ontario, Canada M4N 3M5 (e-mail: david consultant to AdvancePCS, a provider of health 1652 JAMA, April 2, 2003—Vol 289, No. 13 (Reprinted)
2003 American Medical Association. All rights reserved.
DRUG-DRUG INTERACTIONS AMONG ELDERLY PATIENTS ADMITTED FOR DRUG TOXICITY greatest extent to hospital stay46) of hy- tion drugs dispensed to all Ontario resi- poglycemia (International Classification of Diseases, Ninth Revision [ICD-9]47 sible diagnosis of digoxin toxicity (ICD-9 tient and outpatient services, and the On- lemia (ICD-9 code 276.7). The date of date for all analyses. Only the first ad- years or older. These databases have been cause we used a deterministic rather than Control Patients
linkage rate among databases was 100%.
Individual Observation Period
alysis treatments in the preceding year.
tion following his or her 66th birthday.
icity, the end of the study period, death, Setting and Design
and maintained the matching process.
cember 31, 2000) in Ontario, Canada.
older.40 These elderly patients have uni- Exposure to Interacting
Medications
In the glyburide analysis, exposure was trimoxazole. To test the specificity of our Case Patients
of glyburide, we defined case patients as Data Sources
tients but is not known to potentiate the analysis, we identified prescriptions for 2003 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 2, 2003—Vol 289, No. 13 1653
DRUG-DRUG INTERACTIONS AMONG ELDERLY PATIENTS ADMITTED FOR DRUG TOXICITY hibitor analysis, we identified prescrip- sparing diuretics (amiloride, triamterene, est in the 3 years prior to cohort entry.
attributable fraction equals the OR minus Statistical Analysis
than inserted individually (TABLE 1).
scriptions in the week prior to the index hospital admission for drug toxicity, and Sensitivity Analyses
We repeated our analyses with a varietyof modifications to assess the robust-ness of our findings. For patients who ap- Table 1. Covariates Included in the Multivariate Models
Specific Medications
Outcome (Drug)
Potential Predictor
or Conditions*
analysis, adjusting for exposure to other days of the index date. As a final test of Abbreviations: ACE, angiotensin-converting enzyme; CYP, cytochrome P.
*In the primary analysis, adjustment is made for use of any of these medications in 90 days preceding the index date unless otherwise noted. Each predictor category comprises a single term in the multivariate model.
†Other medications include amiodarone, disulfiram, fluconazole, fluvastatin, fluvoxamine, isoniazid, paroxetine, sertra- Primary Analyses
‡Hospital admission for the drug-related toxicity of interest in 3 years preceding cohort entry.
§Other medications include amiodarone, cyclosporine, diltiazem, ketoconazole, nifedipine, propafenone, quinidine, qui- ࿣Medications include combinations of amiloride, triamterene, or spironolactone with a thiazide diuretic.
¶Any prescription for oral hypoglycemic agents or insulin in the 3 years preceding the index date.
a total of 431662 patient-years of therapy.
1654 JAMA, April 2, 2003—Vol 289, No. 13 (Reprinted)
2003 American Medical Association. All rights reserved.
DRUG-DRUG INTERACTIONS AMONG ELDERLY PATIENTS ADMITTED FOR DRUG TOXICITY Table 2. Characteristics of Cases and Controls
Glyburide With
Digoxin With
ACE Inhibitors With
Co-trimoxazole
Clarithromycin
K+-Sparing Diuretics
Controls
Controls
Controls
(n = 43 766)
(n = 1051)
(n = 51 896)
(n = 25 807)
78.2 (72.7-83.3) 80.6 (75.4-85.8) 80.5 (75.3-85.7) 78.4 (72.3-84.5) Residence in long-term care facility, No. (%) Drug first dispensed within 30 days, No. (%) Medication use in preceding year, No. (%) Abbreviations: ACE, angiotensin-converting enzyme; IQR, interquartile range; K+, potassium; NSAIDs, nonsteroidal anti-inflammatory drugs.
The median (interquartile range [IQR])age was 71.6 (67.2-77.3) (7.2) years and Table 3. Association Between Hospital Admission for Hypoglycemia and
Use of Co-trimoxazole in Patients Receiving Glyburide
51% were women (TABLE 2). A total of
No. (%) Exposed
909 patients were admitted to hospitalwith a most responsible diagnosis of hy- Controls
Univariate Odds Ratio
Adjusted Odds Ratio
(n = 43 766)
(95% CI)*
Hospitalization Within 1 Week of Exposure to Second Drug
treated with glyburide for a median (IQR) length of hospital stay for hypoglyce-mia was 4 (2-7) days and 12 patients Hospitalization Within 2 Weeks of Exposure to Second Drug
Hospitalization Within 3 Weeks of Exposure to Second Drug
ing for other factors, cases were about 8 times more likely to have received a pre- Abbreviation: CI, confidence interval.
*Multivariate analysis adjusted for factors in Table 1.
prior to admission (OR, 8.5; 95% CI, 5.8- †Comparable noninteracting drug for comparison.
12.4) (TABLE 3). As expected, we found
no significant association between hy-
preceding week in patients receiving gly- ing digoxin for a total of 513036 patient- lar findings (Table 3). Overall, we esti- mitted to hospital for digoxin toxicity.
mate that at least 3.3% of the hospital ad- 13.6; 95% CI, 8.8-20.8) (TABLE 4). We
2003 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 2, 2003—Vol 289, No. 13 1655
DRUG-DRUG INTERACTIONS AMONG ELDERLY PATIENTS ADMITTED FOR DRUG TOXICITY the hospital because of hyperkalemia.
and other factors (Table 1) that may have estimate that at least 7.8% of the hospi- all, we estimate that at least 2.3% of the ing ACE inhibitors for a total of 1222093 39.9) (TABLE 5). As expected, we found
results. In all cases, the point estimatefrom the bootstrap lies within the 95%CIs of the standard analysis.
Table 4. Association Between Hospital Admission for Digoxin Toxicity and
Use of Clarithromycin in Patients Receiving Digoxin
No. (%) Exposed
and controls whose index date occurredwithin 30 days of commencing therapy Controls
Univariate Odds Ratio
Adjusted Odds Ratio*
with either glyburide, digoxin, or an ACE (n = 1051)
(n = 51 896)
Hospitalization Within 1 Week of Exposure to Second Drug
inhibitor, we found similar results.
started glyburide therapy, 7 (7.6%) alsohad received co-trimoxazole in the pre- Hospitalization Within 2 Weeks of Exposure to Second Drug
Hospitalization Within 3 Weeks of Exposure to Second Drug
Abbreviation: CI, confidence interval.
*Multivariate analysis adjusted for factors in Table 1.
†Comparable noninteracting drug for comparison.
(n=1772), only 20 (1.1%) had receivedclarithromycin. Finally, among cases Table 5. Association Between Hospital Admission for Hyperkalemia and Use of
Potassium-Sparing Diuretics* in Patients Receiving Angiotensin-Converting Enzyme Inhibitors No. (%) Exposed
Controls
Univariate Odds Ratio
Adjusted Odds Ratio
(n = 25 807)
(95% CI)†
Hospitalization Within 1 Week of Exposure to Second Drug
received a potassium-sparing diuretic.
Hospitalization Within 2 Weeks of Exposure to Second Drug
Hospitalization Within 3 Weeks of Exposure to Second Drug
Use of Medications
After Hospital Discharge
Abbreviations: CI, confidence interval; K+, potassium.
*Single-agent potassium-sparing diuretics include amiloride, triamterene, and spironolactone.
†Multivariate analysis adjusted for factors in Table 1.
‡Comparable noninteracting drug for comparison.
1656 JAMA, April 2, 2003—Vol 289, No. 13 (Reprinted)
2003 American Medical Association. All rights reserved.
DRUG-DRUG INTERACTIONS AMONG ELDERLY PATIENTS ADMITTED FOR DRUG TOXICITY tory tests. Furthermore, the high rate of prescriptions in the ensuing 6 months.
times fail at this important task because This is the first study published to date frequent warnings of a trivial nature fa- to use population-based data to study spe- override more significant ones.14,20,22,60 els, renal function, or adherence to medi- cations, and the accuracy of hospital dis- coding tends to attenuate our findings.
ology of other drug-drug interactions.
ing predictable drug-drug interactions in should be aware of these drug-drug inter- sible. In addition, the association with in- controlling for the number of hospital ad- possibly the result of a drug-drug inter- Author Contributions: Study concept and design:
action. The reliable linkage of prescrip- Juurlink, Mamdani, Kopp, Laupacis, Redelmeier.
Acquisition of the data: Kopp.
Analysis and interpretation of the data: Juurlink, Mamdani, Kopp, Laupacis, Redelmeier.
interactions.19 As a result, our findings Drafting of the manuscript: Juurlink, Redelmeier.
reflect only a small fraction of the prob- Critical revision of the manuscript: Juurlink, Mamdani,Kopp, Laupacis, Redelmeier.
Statistical expertise: Juurlink, Mamdani, Redelmeier.
patients, a group that is particularly sus- Obtained funding: Mamdani, Redelmeier.
Administrative, technical, or material support: Kopp, Study supervision: Laupacis, Redelmeier.
Funding/Support: Dr Juurlink was supported by a fel-
conservative estimates that consider only lowship award from the Canadian Institutes of Health Research and the Clinician-Scientist Program of the De- partment of Medicine at the University of Toronto, On-tario. Dr Mamdani was supported by a New Investi- with closer patient monitoring or the use gator award from the New Emerging Teams (NETs) drug toxicity. Finally, ascertainment bias grant of the Canadian Institutes of Health Research, Ot-tawa, Ontario. Dr Laupacis was a Senior Scientist of the Canadian Institutes of Health Research, Ottawa, On- tario. Dr Redelmeier was supported by a career scien-tist award from the Ontario Ministry of Health and the de Sousa chair in trauma at the University of Toronto.
Previous Presentation: Presented at the North Ameri-
can Congress of Clinical Toxicology, Palm Springs, Calif,
of introduction of new drugs and the bur- Acknowledgment: We are grateful to Peter Austin and
of pharmacogenetics.57-59 Reliable, regu- goxin toxicity, and hyperkalemia) are all Deanna Rothwell for providing statistical advice, and wethank Geoff Anderson, Paula Rochon, Gary Naglie, David Babineau, and Ed Etchells for their helpful comments.
2003 American Medical Association. All rights reserved.
(Reprinted) JAMA, April 2, 2003—Vol 289, No. 13 1657
DRUG-DRUG INTERACTIONS AMONG ELDERLY PATIENTS ADMITTED FOR DRUG TOXICITY REFERENCES
1. Lazarou J, Pomeranz BH, Corey PN. Incidence of
ing in community pharmacies and medical practices.
and appendectomy. CMAJ. 2000;162:1421-1424.
adverse drug reactions in hospitalized patients: a meta- Clin Pharmacol Ther. 2001;69:260-265.
44. Mamdani MM, van Walraven C, Bica A, Wil-
analysis of prospective studies. JAMA. 1998;279: 24. Johnson JF, Dobmeier ME. Symptomatic hypo-
liams JI, Naylor CD. Is there an association between glycemia secondary to a glipizide-trimethoprim/ lipid-lowering drugs and cholecystectomy? Am J Med.
2. Einarson TR. Drug-related hospital admissions. Ann
sulfamethoxazole drug interaction. DICP. 1990;24: Pharmacother. 1993;27:832-840.
45. Ray JG, Mamdani M, Tsuyuki RT, Anderson DR,
3. Jha AK, Kuperman GJ, Rittenberg E, Teich JM, Bates
25. Christensen LK, Kristensen M. Drug induced
Yeo EL, Laupacis A. Use of statins and the subse- DW. Identifying hospital admissions due to adverse changes of the blood glucose lowering effect of oral quent development of deep vein thrombosis. Arch In- drug events using a computer-based monitor. Phar- hypoglycemic agents. Acta Diabetol Lat. 1969;(6 suppl tern Med. 2001;161:1405-1410.
macoepidemiol Drug Saf. 2001;10:113-119.
46. Rockhill B, Newman B, Weinberg C. Use and mis-
4. Monahan BP, Ferguson CL, Killeavy ES, Lloyd BK,
26. Potasman I, Bassan H. Hypoglycemia due to in-
use of population attributable fractions. Am J Public Troy J, Cantilena LR Jr. Torsades de pointes occurring teraction between chlorpropamide and cotrimoxa- in association with terfenadine use. JAMA. 1990;264: 47. International Classification of Diseases - Ninth
27. Brian WR. Hypoglycemic agents. In: Levy R, Thum-
Revision, Hospital Edition. 6th ed. Los Angeles, Calif: 5. Wynn RL. Erythromycin and ketoconazole (Ni-
mel K, Trager W, Hansten PD, Eichelbaum M, eds.
Practice Management Information Corp; 2003.
zoral) associated with terfenadine (Seldane)-induced Metabolic Drug Interactions. Philadelphia, Pa: Lip- 48. Schlesselman JJ, Stolley PD. Case-Control Stud-
ventricular arrhythmias. Gen Dent. 1993;41:27-29.
pincott Williams & Wilkins; 2000:529-543.
ies: Design, Conduct, and Analysis. New York, NY: 6. DuBuske LM. Second-generation antihistamines:
28. Kirchheiner J, Brockmoller J, Meineke I, et al. Im-
the risk of ventricular arrhythmias. Clin Ther. 1999; pact of CYP2C9 amino acid polymorphisms on gly- 49. Kleinbaum DG, Kupper LL, Morganstern H. Epi-
buride kinetics and on the insulin and glucose re- demiologic Research. Belmont, Calif: Lifetime Learn- 7. Zechnich AD, Hedges JR, Eiselt-Proteau D, Haxby
sponse in healthy volunteers. Clin Pharmacol Ther.
D. Possible interactions with terfenadine or astemi- 50. Miettinen OS. Proportion of disease caused or pre-
zole. West J Med. 1994;160:321-325.
29. Wen X, Wang JS, Backman JT, Laitila J, Neu-
vented by a given exposure, trait or intervention. Am 8. Wysowski DK, Bacsanyi J. Cisapride and fatal ar-
vonen PJ. Trimethoprim and sulfamethoxazole are se- J Epidemiol. 1974;99:325-332.
rhythmia. N Engl J Med. 1996;335:290-291.
lective inhibitors of CYP2C8 and CYP2C9, respec- 51. Efron B, Tibshirani RJ. An Introduction to the Boot-
9. Michalets EL, Williams CR. Drug interactions with
tively. Drug Metab Dispos. 2002;30:631-635.
strap. New York, NY: Chapman & Hall; 1993.
cisapride: clinical implications. Clin Pharmacokinet.
30. Niemi M, Cascorbi I, Timm R, Kroemer HK, Neu-
52. Schneeweiss S, Seeger JD, Maclure M, Wang PS,
vonen PJ, Kivisto KT. Glyburide and glimepiride phar- Avorn J, Glynn RJ. Performance of comorbidity scores 10. Wysowski DK, Corken A, Gallo-Torres H, Ta-
macokinetics in subjects with different CYP2C9 geno- to control for confounding in epidemiologic studies us- larico L, Rodriguez EM. Postmarketing reports of QT types. Clin Pharmacol Ther. 2002;72:326-332.
ing claims data. Am J Epidemiol. 2001;154:854-864.
prolongation and ventricular arrhythmia in associa- 31. Wakasugi H, Yano I, Ito T, et al. Effect of clarithro-
53. Seymour RM, Routledge PA. Important drug-
tion with cisapride and Food and Drug Administra- mycin on renal excretion of digoxin: interaction with P- drug interactions in the elderly. Drugs Aging. 1998;12: tion regulatory actions. Am J Gastroenterol. 2001;96: glycoprotein. Clin Pharmacol Ther. 1998;64:123-128.
32. Koren G, Woodland C, Ito S. Toxic digoxin-drug
54. Atkin PA, Veitch PC, Veitch EM, Ogle SJ. The epi-
11. Shaoul R, Shahory R, Tamir A, Jaffe M. Compari-
interactions: the major role of renal P-glycoprotein.
demiology of serious adverse drug reactions among son between pediatricians and family practitioners in Vet Hum Toxicol. 1998;40:45-46.
the elderly. Drugs Aging. 1999;14:141-152.
the use of the prokinetic cisapride for gastroesopha- 33. Brown BA, Wallace RJ, Jr, Griffith DE, Warden R.
55. Rosholm JU, Bjerrum L, Hallas J, Worm J, Gram
geal reflux disease in children. Pediatrics. 2002;109: Clarithromycin-associated digoxin toxicity in the el- LF. Polypharmacy and the risk of drug-drug interac- derly. Clin Infect Dis. 1997;24:92-93.
tions among Danish elderly: a prescription database 12. Langdorf MI, Fox JC, Marwah RS, Montague BJ,
34. Gooderham MJ, Bolli P, Fernandez PG. Concomi-
study. Dan Med Bull. 1998;45:210-213.
Hart MM. Physician versus computer knowledge of tant digoxin toxicity and warfarin interaction in a pa- 56. Tamblyn RM, McLeod PJ, Abrahamowicz M,
potential drug interactions in the emergency depart- tient receiving clarithromycin. Ann Pharmacother.
Laprise R. Do too many cooks spoil the broth? mul- ment. Acad Emerg Med. 2000;7:1321-1329.
tiple physician involvement in medical management 13. Weideman RA, Bernstein IH, McKinney WP. Phar-
35. Guerriero SE, Ehrenpreis E, Gallagher KL. Two cases
of elderly patients and potentially inappropriate drug macist recognition of potential drug interactions. Am of clarithromycin-induced digoxin toxicity. Pharma- combinations. CMAJ. 1996;154:1177-1184.
J Health Syst Pharm. 1999;56:1524-1529.
cotherapy. 1997;17:1035-1037.
57. Roden DM, George AL Jr. The genetic basis of
14. Cavuto NJ, Woosley RL, Sale M. Pharmacies and
36. Reardon LC, Macpherson DS. Hyperkalemia in
variability in drug responses. Nat Rev Drug Discov.
prevention of potentially fatal drug interactions. JAMA.
outpatients using angiotensin-converting enzyme in- hibitors: how much should we worry? Arch Intern Med.
58. Guzey C, Spigset O. Genotyping of drug tar-
15. Glassman PA, Simon B, Belperio P, Lanto A. Im-
gets: a method to predict adverse drug reactions? Drug proving recognition of drug interactions: benefits and 37. Kifor I, Moore TJ, Fallo F, et al. Potassium-
barriers to using automated drug alerts. Med Care.
stimulated angiotensin release from superfused adre- 59. Roberts RL, Begg EJ, Joyce PR, Kennedy MA. How
nal capsules and enzymatically dispersed cells of the zona the pharmacogenetics of cytochrome P450 enzymes may 16. Hazlet TK, Lee TA, Hansten PD, Horn JR. Perfor-
glomerulosa. Endocrinology. 1991;129:823-831.
affect prescribing. N Z Med J. 2002;115:137-140.
mance of community pharmacy drug interaction soft- 38. Schepkens H, Vanholder R, Billiouw JM, Lameire
60. Jones JK, Fife D, Curkendall S, Goehring E Jr, Guo
ware. J Am Pharm Assoc (Wash). 2001;41:200-204.
N. Life-threatening hyperkalemia during combined JJ, Shannon M. Coprescribing and codispensing of 17. Del Fiol G, Rocha BH, Kuperman GJ, Bates DW,
therapy with angiotensin-converting enzyme inhibi- cisapride and contraindicated drugs. JAMA. 2001; Nohama P. Comparison of two knowledge bases on tors and spironolactone: an analysis of 25 cases. Am the detection of drug-drug interactions. Proc AMIA 61. Pitt B, Zannad F, Remme WJ, et al. The effect of
39. Ahuja TS, Freeman D Jr, Mahnken JD, Agraha-
spironolactone on morbidity and mortality in pa- 18. Shapiro LE, Shear NH. Drug-drug interactions: how
rkar M, Siddiqui M, Memon A. Predictors of the de- tients with severe heart failure: Randomized Aldac- scared should we be? CMAJ. 1999;161:1266-1267.
velopment of hyperkalemia in patients using angio- tone Evaluation Study Investigators. N Engl J Med.
19. Peterson JF, Bates DW. Preventable medication er-
tensin-converting enzyme inhibitors. Am J Nephrol.
rors: identifying and eliminating serious drug interac- 62. Heeley E, Riley J, Layton D, Wilton LV, Shakir SA.
tions. J Am Pharm Assoc (Wash). 2001;41:159-160.
40. Statistics Canada. Population by age group—
Prescription-event monitoring and reporting of ad- 20. Hansten PD, Horn JR, Hazlet TK. ORCA: OpeRa-
Statistics Canada. Available at: http://www.statcan.ca verse drug reactions. Lancet. 2001;358:1872-1873.
tional ClassificAtion of drug interactions. J Am Pharm /english/Pgdb/demo31b.htm. Accessed March 10, 63. Alvarez-Requejo A, Carvajal A, Begaud B, Mo-
Assoc (Wash). 2001;41:161-165.
ride Y, Vega T, Arias LH. Under-reporting of adverse 21. Leape LL, Brennan TA, Laird N, et al. The nature
41. Bell CM, Redelmeier DA. Mortality among pa-
drug reactions: estimate based on a spontaneous re- of adverse events in hospitalized patients: results of tients admitted to hospitals on weekends as compared porting scheme and a sentinel system. Eur J Clin Phar- the Harvard Medical Practice Study II. N Engl J Med.
with weekdays. N Engl J Med. 2001;345:663-668.
42. Crighton EJ, Mamdani MM, Upshur RE. A popu-
64. Moride Y, Haramburu F, Requejo AA, Begaud B.
22. Bates DW, Cullen DJ, Laird N, et al. Incidence of
lation based time series analysis of asthma hospitali- Under-reporting of adverse drug reactions in general adverse drug events and potential adverse drug events: sations in Ontario, Canada: 1988 to 2000. BMC Health practice. Br J Clin Pharmacol. 1997;43:177-181.
implications for prevention: ADE Prevention Study 65. Lehmann DF. Observation and experiment on the
43. Mamdani MM, Tu K, van Walraven C, Austin PC,
cusp of collaboration: a parallel examination of clini- 23. Halkin H, Katzir I, Kurman I, Jan J, Malkin BB. Pre-
Naylor CD. Postmenopausal estrogen replacement cal pharmacology and pharmacoepidemiology. J Clin venting drug interactions by online prescription screen- therapy and increased rates of cholecystectomy 1658 JAMA, April 2, 2003—Vol 289, No. 13 (Reprinted)
2003 American Medical Association. All rights reserved.

Source: http://biostat.jhsph.edu/courses/bio624/misc/JAMA%20_Example1652.pdf