Cox proportional hazards regression models are frequently used to determine the association between exposure and time-to-event outcomes in both randomized controlled trials and in observational cohort studies. The resultant hazard ratio is a relative measure of effect that provides limited clinical information.
A method is described for deriving absolute reductions in the risk of an event occurring within a given duration of follow-up time from a Cox regression model. The associated number needed to treat can be derived from this quantity. The method involves determining the probability of the outcome occurring within the specified duration of follow-up if each subject in the cohort was treated and if each subject was untreated, based on the covariates in the regression model. These probabilities are then averaged across the study population to determine the average probability of the occurrence of an event within a specific duration of follow-up in the population if all subjects were treated and if all subjects were untreated.
Risk differences and numbers needed to treat.
Absolute measures of treatment effect can be derived in prospective studies when Cox regression is used to adjust for possible imbalance in prognostically important baseline covariates.
Higher blood pressure in acute heart failure has been associated with improved survival; however, the relationship between blood pressure and survival in stabilized patients at hospital discharge has not been established.
In 7448 patients with heart failure (75.2+/-11.5 years; 49.9% men) discharged from the hospital in Ontario, Canada, we examined the association of systolic blood pressure (SBP) and diastolic blood pressure with long-term survival. Parametric survival analysis was performed, and survival time ratios were determined according to discharge blood pressure group. A total of 25 427 person-years of follow-up were examined. In those with left ventricular ejection fraction or =160 mm Hg, respectively. In those with left ventricular ejection fraction >40%, survival time ratios were 0.69 (95% CI, 0.51 to 0.93), 0.83 (95% CI, 0.71 to 0.99), 0.95 (95% CI, 0.80 to 1.14), and 0.76 (95% CI, 0.61 to 0.95) for discharge SBPs or =160 mm Hg, respectively.
In this long-term population-based study of patients with heart failure, the association of discharge SBP with mortality followed a U-shaped distribution. Survival was shortened in those with reduced or increased values of discharge SBP.
It is not known whether subspecialty care by cardiologists improves outcomes in heart failure patients from the community over care by other physicians.
Using administrative data, we monitored 38 702 consecutive patients with first-time hospitalization for heart failure in Ontario, Canada, between April 1994 and March 1996 and examined differences in processes of care and clinical outcomes between patients attended by physicians of different disciplines. We found that patients attended by cardiologists had lower 1-year risk-adjusted mortality than those attended by general internists, family practitioners, and other physicians (28.5% versus 31.7%, 34.9%, and 35.9%, respectively; all pairwise comparisons, P
Investigators in clinical research are often interested in determining the association between patient characteristics and cost of medical or surgical treatment. However, there is no uniformly agreed upon regression model with which to analyse cost data. The objective of the current study was to compare the performance of linear regression, linear regression with log-transformed cost, generalized linear models with Poisson, negative binomial and gamma distributions, median regression, and proportional hazards models for analysing costs in a cohort of patients undergoing CABG surgery. The study was performed on data comprising 1959 patients who underwent CABG surgery in Calgary, Alberta, between June 1994 and March 1998. Ten of 21 patient characteristics were significantly associated with cost of surgery in all seven models. Eight variables were not significantly associated with cost of surgery in all seven models. Using mean squared prediction error as a loss function, proportional hazards regression and the three generalized linear models were best able to predict cost in independent validation data. Using mean absolute error, linear regression with log-transformed cost, proportional hazards regression, and median regression to predict median cost, were best able to predict cost in independent validation data. Since the models demonstrated good consistency in identifying factors associated with increased cost of CABG surgery, any of the seven models can be used for identifying factors associated with increased cost of surgery. However, the magnitude of, and the interpretation of, the coefficients vary across models. Researchers are encouraged to consider a variety of candidate models, including those better known in the econometrics literature, rather than begin data analysis with one regression model selected a priori. The final choice of regression model should be made after a careful assessment of how best to assess predictive ability and should be tailored to the particular data in question.
Although approximately one third of patients with heart failure (HF) visiting the emergency department (ED) are discharged home, little is known about their care and outcomes.
We examined the acute care and early outcomes of patients with HF who visited an ED and were discharged without hospital admission in Ontario, Canada, from April 2004 to March 2007. Among 50 816 patients (age, 76.4+/-11.6 years; 49.4% men) visiting an ED for HF, 16 094 (31.7%) were discharged without hospital admission. A total of 4.0% died within 30 days from admission, and 1.3% died within 7 days of discharge from the ED. Although multiple (>or=2) previous HF admissions (odds ratio [OR], 1.64; 95% CI, 1.14 to 2.31), valvular heart disease (OR, 1.37; 95% CI, 1.00 to 1.84), peripheral vascular disease (OR, 1.41; 95% CI, 1.00 to 1.93), and respiratory disease (OR, 1.33; 95% CI, 1.08 to 1.63) increased the risk of 30-day death among those discharged from the ED, presence of these conditions did not increase the likelihood of admission. Patients were more likely to be admitted if they were older (OR, 1.08; 95% CI, 1.06 to 1.10 per decade), arrived by ambulance (OR, 2.02; 95% CI, 1.93 to 2.12), had a higher triage acuity score (OR, 4.12; 95% CI, 3.84 to 4.42), or received resuscitation in the ED (OR, 2.85; 95% CI, 2.68 to 3.04). In those with comparable predicted risks of death, subsequent 90-day mortality rates were higher among discharged than admitted patients (11.9% versus 9.5%; log-rank P=0.016).
Patients with HF who are discharged from the ED have substantial risks of early death, which, in some cases, may exceed that of hospitalized patients.
To determine if the delivery of hospital discharge summaries to follow-up physicians decreases the risk of hospital readmission.
Eight hundred eighty-eight patients discharged from a single hospital following treatment for an acute medical illness.
Teaching hospital in a universal health-care system.
We determined the date that each patient's discharge summary was printed and the physicians to whom it was sent. Summary receipt was confirmed by survey and phoning each physician's office. Each patient's hospital chart was reviewed to determine their acute and chronic medical conditions as well as their course in hospital. Using population-based administrative databases, all post-hospitalization visits were identified. For each of these visits, we determined whether the summary was available.
Time to nonelective hospital readmission during 3 months following discharge.
The discharge summary was available for only 568 of 4,639 outpatient visits (12.2%). Overall, 240 (27.0%) of patients were urgently readmitted to hospital. After adjusting for significant patient and hospitalization factors, we found a trend toward a decreased risk of readmission for patients who were seen in follow-up by a physician who had received a summary (relative risk 0.74, 95% confidence interval 0.50 to 1.11).
The risk of rehospitalization may decrease when patients are assessed following discharge by physicians who have received the discharge summary. Further research is required to determine if better continuity of patient information improves patient outcomes.
Cites: J Gen Intern Med. 1991 May-Jun;6(3):223-82066826
Cites: Ann Intern Med. 1999 Feb 16;130(4 Pt 2):338-4210068402
Cites: J Gen Intern Med. 1991 Sep-Oct;6(5):389-931744751
Cites: Health Serv Res. 1992 Apr;27(1):81-1011563955
Cites: Inquiry. 1993 Spring;30(1):95-1038454320
Cites: J Am Geriatr Soc. 1993 Aug;41(8):811-78340558
Cites: Inquiry. 1994 Summer;31(2):163-728021022
Cites: Arch Intern Med. 1994 Aug 8;154(15):1721-98042889
Cites: Aust Fam Physician. 1994 Jul;23(7):1320-1, 1324-5, 13288060277
Cites: Ann Intern Med. 1995 Mar 15;122(6):415-217856989
Cites: CMAJ. 1995 May 1;152(9):1423-337728691
Cites: CMAJ. 1995 May 1;152(9):1437-427728692
Cites: Arch Fam Med. 1995 May;4(5):403-97742962
Cites: Br J Surg. 1996 Jun;83(6):788-908696741
Cites: Ann Intern Med. 1999 Feb 16;130(4 Pt 2):343-910068403
Cites: Ann Intern Med. 1999 Feb 16;130(4 Pt 2):368-7210068408
Cites: Ann Intern Med. 1999 Feb 16;130(4 Pt 2):373-510068409
Cites: Ann Intern Med. 1999 Feb 16;130(4 Pt 2):376-8110068410
Cites: Ann Intern Med. 1999 Feb 16;130(4 Pt 1):326-710068394
Cites: CMAJ. 1999 Feb 9;160(3):319-2610065073
Cites: CMAJ. 1999 Feb 9;160(3):345-610065076
Cites: Am Heart J. 1999 May;137(5):919-2710220642
Cites: Med Care. 1999 May;37(5):490-50110335751
Cites: J Gen Intern Med. 1996 Nov;11(11):684-89120655
Recent debate has emerged regarding the cardiovascular safety of selective cyclooxygenase 2 inhibitors and the possible cardioprotective effect of naproxen sodium. We compared the rates of acute myocardial infarction (AMI) among elderly patients dispensed selective cyclooxygenase 2 inhibitors, naproxen, and nonselective nonnaproxen nonsteroidal anti-inflammatory drugs (NSAIDs).
We conducted a population-based retrospective cohort study using administrative health care data from Ontario, Canada, from April 1, 1998, to March 31, 2001. We identified NSAID-naive cohorts of subjects aged 66 years and older in whom treatment was initiated with celecoxib (n = 15 271), rofecoxib (n = 12 156), naproxen (n = 5669), and nonnaproxen nonselective NSAIDs (n = 33 868), along with a randomly selected control cohort not exposed to NSAIDs (n = 100 000). Multivariate Cox proportional hazards models were used to compare AMI rates between study drug groups while controlling for potential confounders.
Relative to control subjects, the multivariate model showed no significant differences in AMI risk for new users of celecoxib (adjusted rate ratio [aRR], 0.9; 95% confidence interval [CI], 0.7-1.2), rofecoxib (aRR, 1.0; 95% CI, 0.8-1.4), naproxen (aRR, 1.0; 95% CI, 0.6-1.7), or nonnaproxen nonselective NSAIDs (aRR, 1.2; 95% CI, 0.9-1.4).
The findings of this observational study suggest no increase in the short-term risk of AMI among users of selective cyclooxygenase 2 inhibitors as commonly used in clinical practice. Furthermore, the findings do not support a short-term reduced risk of AMI with naproxen.
There are few comparative data regarding long-term nonfatal outcomes for women versus men after coronary artery bypass grafting (CABG). This study compares gender differences in cardiac events in a population of hospital survivors up to 11 years after isolated CABG surgery in Ontario, Canada.
A population-based cohort study (n=68,774 patients, 15,043 women) between September 1, 1991, and April 1, 2002, was assembled with linked clinical and administrative databases. Cox modeling and propensity score matching were used to compare death, cardiac readmission (angina, heart failure, myocardial infarction), repeat revascularization (angioplasty or CABG), and stroke readmission between men and women. Women were older (65+/-17 versus 62+/-13 years), more likely to present with urgent or emergent status (64% versus 56%), and less likely to receive arterial grafts (70% versus 78%). Women had a higher rate of cardiac readmission in the first year after surgery (hazard ratio [HR] of 1.5, 95% confidence interval [CI] 1.36 to 1.56), and this increased risk persisted after 1 year (HR 1.2, 95% CI 1.14 to 1.31). This was primarily due to readmissions for unstable angina (HR 1.3, 95% CI 1.24 to 1.38) and congestive heart failure (HR 1.1, 95% CI 1.06 to 1.21). Propensity-matched women had similar rates of death (HR 0.9, 95% CI 0.83 to 0.98) and repeat revascularization (HR 1.0, 95% CI 0.91 to 1.06).
Women have a more complex clinical preoperative presentation and are more likely to be readmitted with unstable angina and congestive heart failure after CABG but experience survival similar to those seen in men. Gender differences in outcomes may be improved through durable revascularization strategies and close postoperative follow-up care targeted to women.
Incidental abdominal aortic aneurysms (AAAs) are identified when the abdomen is imaged for other reasons. These are common, and many undergo incomplete radiological monitoring. The association between monitoring completeness and population-based outcomes has not been studied.
A cohort of incidental AAAs (defined as previously unidentified aortic enlargement exceeding 3 cm found on an imaging study done for another reason) was linked to population-based data. Patients were followed to elective AAA repair, AAA rupture, death, or March 31, 2009. Monitoring completeness was gauged as the sequential number of months without a recommended abdominal scan. Its association with time to elective AAA repair and time to death was measured using a multivariable Cox regression model adjusting for other important covariates.
We identified 191 incidental AAAs between 1996 and 2004 (median diameter of 3.5 cm [range, 3.0-5.3 cm], median follow up of 4.4 years [range, 0.6-12.7 years]). During the study, patients spent a median of 19.4% of their time with incomplete AAA monitoring (interquartile range [IQR] 0.3%-44%); 56 patients (29.3%) had no follow-up imaging of their aneurysm. Nineteen patients (10.0%; 2.0% per year) underwent elective AAA repair, and 79 patients (37.7%; 7.6% per year) died. Independent of important covariates, people were significantly less likely to undergo elective repair (hazard ratio [HR], 0.03) and significantly more likely to die (HR, 2.99) if their AAA went without radiological monitoring for 1 year.
Incomplete incidental AAA radiological monitoring was significantly associated with a decreased risk of elective AAA repair and an increased risk of death. While uncontrolled confounding might explain part of these associations, clinicians should ensure that radiological monitoring of AAAs is complete in appropriate patients.
Hospital readmissions for heart failure (HF) contribute to increased morbidity and resource burden. Predictors of hospitalization and patterns of cardiovascular events over the lifetime of patients with HF have not been elucidated.
We examined recurrent hospitalizations, cardiovascular events, and survival among newly discharged (April 1999-March 2001) patients with HF in the Enhanced Feedback For Effective Cardiac Treatment phase 1 study. During 10-year follow-up, we examined all new cardiovascular hospitalizations and selected predictors of readmission. Among 8543 patients (mean age, 77.4±10.5 years; 51.6% women) followed for 22 567 person-years, 60.7% had ischemic etiology, and 67.3% had HF with reduced ejection fraction (left ventricular ejection fraction =45% versus >45% [HF with preserved ejection fraction]). Overall, 10-year mortality was 98.8%, with 35 966 hospital readmissions occurring over the lifetime of the cohort. Adjusted hazards ratios (HRs) for first cardiovascular hospitalization were 1.36 for ischemic HF (95% CI, 1.28-1.44; P