Portal de Periódicos da CAPES

The Dialysis Outcomes and Practice Patterns Study (DOPPS): An international hemodialysis study

2000; Elsevier BV; Volume: 57; Linguagem: Inglês

10.1046/j.1523-1755.2000.07413.x

1523-1755

Eric W. Young, David A. Goodkin, Donna Mapes, Friedrich K. Port, Marcia Keen, Kenneth Chen, Bradley L. Maroni, Robert A. Wolfe, Philip J. Held,

Chronic Kidney Disease and Diabetes

The Dialysis Outcomes and Practice Patterns Study (DOPPS): An international hemodialysis study. The Dialysis Outcomes and Practice Patterns Study (DOPPS) is a prospective, longitudinal, observational study of hemodialysis patients and facilities in seven countries with large populations of dialysis patients: France, Germany, Italy, Japan, Spain, the United Kingdom, and the United States. This paper describes the study design, analytic methods, and preliminary findings of the DOPPS. The goal of the study is to determine which practice patterns are associated with the best patient outcomes, with adjustment for a wide range of patient case-mix characteristics. The primary outcomes of interest are mortality, hospitalization, quality of life, and vascular access events. The facility sample from the seven countries consists of 327 hemodialysis centers in which 24,392 patients were treated when the study began. A random sample of 10,332 patients has been selected thus far for more detailed longitudinal data collection. Departing patients are replaced during the study using random selection. A study coordinator at each dialysis facility collects baseline and longitudinal patient data. Patients are asked to complete a questionnaire that addresses quality of life on a yearly basis. The medical director and nurse manager in each facility complete a practice pattern questionnaire. Preliminary data are presented concerning the sample facilities and the census of patients treated in each facility at the start of the study. Dialysis facilities vary widely in size and type (freestanding vs. institutionally-based) across countries. Variation is also seen in patient age, sex distribution, and diabetes mellitus as the attributed cause of end-stage renal disease (ESRD). At this early phase, the DOPPS has proved to be technically feasible and has revealed basic differences in hemodialysis facilities and patients across the seven participating countries. The Dialysis Outcomes and Practice Patterns Study (DOPPS): An international hemodialysis study. The Dialysis Outcomes and Practice Patterns Study (DOPPS) is a prospective, longitudinal, observational study of hemodialysis patients and facilities in seven countries with large populations of dialysis patients: France, Germany, Italy, Japan, Spain, the United Kingdom, and the United States. This paper describes the study design, analytic methods, and preliminary findings of the DOPPS. The goal of the study is to determine which practice patterns are associated with the best patient outcomes, with adjustment for a wide range of patient case-mix characteristics. The primary outcomes of interest are mortality, hospitalization, quality of life, and vascular access events. The facility sample from the seven countries consists of 327 hemodialysis centers in which 24,392 patients were treated when the study began. A random sample of 10,332 patients has been selected thus far for more detailed longitudinal data collection. Departing patients are replaced during the study using random selection. A study coordinator at each dialysis facility collects baseline and longitudinal patient data. Patients are asked to complete a questionnaire that addresses quality of life on a yearly basis. The medical director and nurse manager in each facility complete a practice pattern questionnaire. Preliminary data are presented concerning the sample facilities and the census of patients treated in each facility at the start of the study. Dialysis facilities vary widely in size and type (freestanding vs. institutionally-based) across countries. Variation is also seen in patient age, sex distribution, and diabetes mellitus as the attributed cause of end-stage renal disease (ESRD). At this early phase, the DOPPS has proved to be technically feasible and has revealed basic differences in hemodialysis facilities and patients across the seven participating countries. Dialysis therapy ameliorates many of the clinical manifestations of renal failure and postpones otherwise imminent death. Despite this undeniable success, hemodialysis patients have higher mortality and hospitalization rates and lower quality of life than the general population. Also, patients must contend with unique treatment complications, such as vascular access failure, at considerable expense and morbidity. The Dialysis Outcomes and Practice Patterns Study (DOPPS) is a prospective, international, observational study of hemodialysis practices and associated outcomes. The primary goal of DOPPS is to improve the understanding of dialysis practices that are associated with better outcomes for patients. The primary study endpoints are mortality, hospitalization, quality of life, and vascular access outcomes. DOPPS is predicated on several observations. First, although the overall mortality rate is exceedingly high among dialysis patients, outcomes vary substantially across facilities and countries1McClellan W.M. Flanders W.D. Gutman R.A. Variable mortality rates among dialysis treatment centers.Ann Intern Med. 1992; 117: 332-336Crossref PubMed Scopus (66) Google Scholar, 2Hulbert-Shearon T.E. Loos E. Ashby V.B. Port F.K. Wolfe R.A. USRDS 1999 Unit-Specific Reports for Dialysis Patients: A Summary. University of Michigan, Ann Arbor1999Google Scholar, 3Held P.J. Brunner F. Odaka M. Garcia J.R. Port F.K. Gaylin D.S. Five-year survival for end-stage renal disease patients in the United States, Europe, and Japan, 1982–87.Am J Kidney Dis. 1990; 15: 451-457Abstract Full Text PDF PubMed Scopus (259) Google Scholar, 4Owen W.F. Lew N.L. Liu Y. Lowrie E.G. Lazarus J.M. The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis.N Engl J Med. 1993; 329: 1001-1006Crossref PubMed Scopus (1165) Google Scholar. For example, a fivefold variation in crude mortality was reported across facilities in the United States and adjusted mortality indicators show comparable variability1McClellan W.M. Flanders W.D. Gutman R.A. Variable mortality rates among dialysis treatment centers.Ann Intern Med. 1992; 117: 332-336Crossref PubMed Scopus (66) Google Scholar,2Hulbert-Shearon T.E. Loos E. Ashby V.B. Port F.K. Wolfe R.A. USRDS 1999 Unit-Specific Reports for Dialysis Patients: A Summary. University of Michigan, Ann Arbor1999Google Scholar. Also, the reported 5-year mortality rates for end-stage renal disease (ESRD) patients in Europe and Japan are 20–35% lower than those reported for patients in the United States, even with adjustment for age, sex, and diabetic status3Held P.J. Brunner F. Odaka M. Garcia J.R. Port F.K. Gaylin D.S. Five-year survival for end-stage renal disease patients in the United States, Europe, and Japan, 1982–87.Am J Kidney Dis. 1990; 15: 451-457Abstract Full Text PDF PubMed Scopus (259) Google Scholar. The observed variation in mortality across centers and countries raises the strong possibility that differing treatment practices may contribute to the variation in outcomes. Second, dialysis outcomes can be modified by changes in dialysis practice. For example, several studies have shown that improved patient survival is associated with higher dialysis doses and use of different types of dialysis membranes4Owen W.F. Lew N.L. Liu Y. Lowrie E.G. Lazarus J.M. The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis.N Engl J Med. 1993; 329: 1001-1006Crossref PubMed Scopus (1165) Google Scholar, 5Held P.J. Port F.K. Wolfe R.A. Stannard D.C. Carroll C.E. Daugiridas J.T. Greer J.W. Hakim R.M. The dose of hemodialysis and patient mortality.Kidney Int. 1996; 50: 550-556Abstract Full Text PDF PubMed Scopus (466) Google Scholar, 6Hakim R.M. Held P.J. Stannard D.C. Wolfe R.A. Port F.K. Daugirdas J.T. Agodoa L. Effect of the dialysis membrane on mortality of chronic hemodialysis patients.Kidney Int. 1996; 50: 566-570Abstract Full Text PDF PubMed Scopus (176) Google Scholar. Finally, observational studies have proven to be an efficient means for discovering associations between treatment patterns and outcomes4Owen W.F. Lew N.L. Liu Y. Lowrie E.G. Lazarus J.M. The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis.N Engl J Med. 1993; 329: 1001-1006Crossref PubMed Scopus (1165) Google Scholar, 5Held P.J. Port F.K. Wolfe R.A. Stannard D.C. Carroll C.E. Daugiridas J.T. Greer J.W. Hakim R.M. The dose of hemodialysis and patient mortality.Kidney Int. 1996; 50: 550-556Abstract Full Text PDF PubMed Scopus (466) Google Scholar, 7Churchill D.N. Taylor D.W. Cook R.J. Laplante P. Barre P. Cartier P. Fay W.P. Goldstein M.B. Jindal K. Mandin H. McKenzie J.K. Muirhead N. Parfrey P.S. Posen G.A. Slaughter D. Ulan R.A. Werb R. Canadian hemodialysis morbidity study.Am J Kidney Dis. 1992; 19: 214-234Abstract Full Text PDF PubMed Scopus (486) Google Scholar, 8Greenfield S. Sullivan L. Sillman R.A. Dukes K. Kaplan S.H. Principles and practice of case mix adjustment: applications to end-stage renal disease.Am J Kidney Dis. 1994; 24: 298-307Abstract Full Text PDF PubMed Scopus (41) Google Scholar, 9Held P.J. Wolfe R.A. Gaylin D.S. Port F.K. Levin N.W. Turenne M.N. Analysis of the association of dialyzer reuse practices and patient outcomes.Am J Kidney Dis. 1994; 23: 692-708Abstract Full Text PDF PubMed Scopus (89) Google Scholar. Findings from large, well-designed, nationally representative observational studies have prompted changes in the national practice of dialysis and provided the impetus for important clinical trials10National Kidney Foundation Dialysis Outcomes Quality Initiative Clinical Practice Guidelines.Am J Kidney Dis. 1997; 30: S1-S240Google Scholar,11Eknoyan G. Levey A.S. Beck G.J. Agadoa L.Y. Daugirdas J.T. Kusek J.W. Levin N.W. Schulman G. The Hemodialysis (HEMO) Study: Rationale for selection of interventions.Semin Dial. 1996; 9: 24-33Crossref Google Scholar. DOPPS was designed to address some of the limitations of international registries, which often rely on voluntary reporting and collect little information about practice patterns and individual patient characteristics. DOPPS also addresses the limitations of studies involving single facilities or convenience samples of facilities by studying a representative sample of dialysis centers and patients. DOPPS was designed to study inter-facility variation in hemodialysis practices and outcomes in the hopes of identifying important, potentially causal associations. Variation in practice patterns was achieved through a sample design that included dialysis facilities from seven developed countries with large ESRD populations: France, Germany, Italy, Japan, Spain, the United Kingdom, and the United States. The study features longitudinal collection of dialysis facility practices, patient outcomes, and patient demographics, comorbidities, and laboratory values. The study is unique in the breadth and detail of the information collected regarding demographic characteristics, comorbid conditions, clinical outcomes, and treatment practices for each patient and center. Our intent is to clarify risk factors for deleterious patient outcomes while accounting for the confounding effects of patient case-mix. This report describes the study design and delineates the baseline characteristics of the dialysis facilities and patients. DOPPS is a prospective study of hemodialysis patients and facilities in seven countries. The countries were selected on the basis of geographic diversity, variation in practices and outcomes, and relatively large numbers of ESRD patients. A nationally representative sample of dialysis facilities has been enrolled in each country. A random sample of hemodialysis patients is selected within each participating center. Practice patterns are determined at the facility and patient levels. Demographic, laboratory, comorbidity, and outcome data are ascertained at the patient level. The basic study design and study instruments are shared across all countries with minor local modifications as necessary. Institutional review boards approved the study in each country or facility, as required. Informed patient consent was obtained in accordance with the requirements of each country, review board, and dialysis center. Data collection is performed in a fashion that maintains patient anonymity at the coordinating center. In the United States, a stratified random sample of chronic hemodialysis facilities was selected to achieve variation in practice patterns and outcomes. The initial sampling frame consisted of a random subsample of a listing of dialysis facilities published by the Health Care Financing Administration (HCFA). For each dialysis facility, a measure of mortality was estimated using the adjusted mortality ratio (AMR) for the year 1996, based on publicly available measures of crude mortality and mean age and percentage of diabetic patients in each facility12Loos E.A. Wolfe R.A. Hulbert-Shearon T.E. Held P.J. Quantitative measures of mortality in dialysis units: an alternative to the standardized mortality ratio.J Am Soc Nephrol. 1997; 8: 202AGoogle Scholar. The AMR estimates the standardized mortality ratio (SMR), which is the ratio of observed to expected deaths in a facility where expected deaths are based on the age, race, sex, and diabetic status of each patient treated in the center13Wolfe R.A. The standardized mortality ratio revisited: improvements, innovations, and limitations.Am J Kidney Dis. 1994; 24: 290-297Abstract Full Text PDF PubMed Scopus (59) Google Scholar. The AMR is based on average rather than individual patient characteristics. A simple random sample of 97 dialysis facilities was initially selected without regard to the AMR. In order to augment the representation of facilities with extremes in outcomes (and potentially in practice patterns), a purposive sample of 31 facilities was drawn from the upper tail and 33 from the lower tail of the AMR distribution. A representative description of U.S. dialysis facilities can be obtained using the random sample or the overall sample weighted by the probability of facility selection. In Europe, the sample is composed of 20 dialysis centers from each of the five participating countries for a total of 100 facilities. Facilities were sampled from all dialysis facilities in each country, obtained from national sources with the assistance of the country investigators. Within each country, the sample was proportionately stratified by geographic region and facility type in order to assure that the facilities were representative Table 1. As the number of geographic regions generally exceeded the sample size, sample stratification was achieved by setting a maximum quota for each region. The facility type strata were determined by the conventions used in each country (e.g., center vs. satellite centers, described in Table 1). In Japan, 66 dialysis facilities were enrolled from a national list of hemodialysis facilities. The sample was stratified by geographic region (prefecture) and facility type Table 1.Table 1Summary of hemodialysis facilities in sampling frame for each country, 1996–1999CountryTotal facilitiesTotal HD patientsNo. of geographic regionsaNumber of regions in each country sample frame used in sampling plan. U.S. sample was not stratified by geographic region.No. of facility typesbNumber of facility types used in sampling plan. U.S. sample was not stratified by facility type although facility strata in common use are shown for comparison with other countries.Facility typesbNumber of facility types used in sampling plan. U.S. sample was not stratified by facility type although facility strata in common use are shown for comparison with other countries.France31620,106224General, private, university, associationGermany79043,747cData obtained directly from European Dialysis and Transplant Association (EDTA) for 1995103Clinics (medical centers), nonprofit free-standing, private practiceItaly59930,963cData obtained directly from European Dialysis and Transplant Association (EDTA) for 1995192Public, privateJapan2653158,222dData from Japanese Society for Dialysis Therapy for 199717472Hospital, clinicSpain47615,427cData obtained directly from European Dialysis and Transplant Association (EDTA) for 1995832Academic, non-academicUnited Kingdom1467,528dData from Japanese Society for Dialysis Therapy for 199717112Center, satelliteUnited States2894159,349eData from United States Renal Data System (USRDS) for 19951613Free-standing profit, free-standing non-profit, hospitalHD = hemodialysis, No. = numbera Number of regions in each country sample frame used in sampling plan. U.S. sample was not stratified by geographic region.b Number of facility types used in sampling plan. U.S. sample was not stratified by facility type although facility strata in common use are shown for comparison with other countries.c Data obtained directly from European Dialysis and Transplant Association (EDTA) for 1995d Data from Japanese Society for Dialysis Therapy for 199717Japanese Society for Dialysis Therapy An Overview of Regular Dialysis Treatment in Japan (as of December 31, 1997). 1997Google Scholare Data from United States Renal Data System (USRDS) for 199516US Renal Data System USRDS 1997 Annual Data Report. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD1997Google Scholar Open table in a new tab HD = hemodialysis, No. = number Facilities treating fewer than 20 chronic hemodialysis patients in the US and fewer than 25 elsewhere were excluded for reasons of study efficiency (the minimum size was increased for Europe and Japan because fewer facilities were enrolled than in the United States). This restriction led to the exclusion of fewer than 5% of all hemodialysis patients in each country. Facilities treating less than the threshold number of patients after recruitment (because the patient count declined after facility enrollment) were retained in the study. At the start of the project, the study coordinator in each participating facility listed the census of prevalent in-center hemodialysis patients older than 17 years. This census listing includes basic patient information such as age, race, sex, and the cause of ESRD. At regular intervals of approximately every four months, the census is updated to indicate all new and departed patients since the last census update. The date and reason for each departure are entered on the census form. Within each participating facility, the census listing was used to select a random sample of 20–40 patients, varying according to the size of the facility. Detailed longitudinal data collection is performed for this representative sample of patients. Departed patients are replaced approximately every four months, using random selection from the patients entering the dialysis facility during the interval. A study coordinator in each participating dialysis center performs data collection. In addition, specific questionnaires are completed by sampled patients, the medical director, and the nurse manager of each participating facility. The cumulative census form provides basic data about all hemodialysis patients treated in each facility (see above). The study coordinator completes a detailed medical questionnaire for each patient selected for the sample. Medical questionnaire information is largely abstracted from the medical record, supplemented by personal knowledge of the patients. The medical questionnaire addresses a variety of areas including ESRD history, medical and psychosocial history, dialysis prescription, laboratory data, and prescribed medications at the time of study enrollment Table 2. The study coordinator completes an interval summary approximately every four months for each sampled patient. The interval summary updates laboratory data, dialysis prescription, medication use and the interval occurrence of hospitalizations, outpatient events and medical interventions, vascular access events, and departures Table 2. Patients are asked to complete a questionnaire that includes the Kidney Disease Quality of Life survey (KDQOL™)14Hays R.D. Kallich J.D. Mapes D.L. Coons S.J. Carter W.B. Development of the Kidney Disease Quality of Life (KDQOLTM) instrument.Quality Life Res. 1994; 3: 329-338Crossref PubMed Scopus (756) Google Scholar and modules concerning pre-ESRD care, economic aspects of ESRD, employment and rehabilitation. Patients repeat the KDQOL™ survey each year.Table 2Patient level data collection itemsCategoryIllustrative data itemsMQaMedical Questionnaire (MQ) provides baseline information at time of patient enrollment in DOPPSISbInterval Summary (IS) provides follow-up information for each 4-month study intervalBackground and demographic informationAge, sex, race, ESRD historyX—Insurance coveragePrimary and secondary payerXXMedical historyCause(s) of ESRD, tobacco use, heart disease, cerebrovascular disease, peripheral vascular disease, diabetes mellitus, pulmonary disease, neurologic disease, psychiatric disease, musculoskeletal disease, gastrointestinal and hepatic diseases, cancer, eye disease, hepatitis, HIV, family historyX—Pre-ESRD treatmentNephrologist visit, vascular accessX—Vital signsBlood pressure, weightXXDialysis prescriptionTime, dialyzer, dialysate composition, blood flowXXPsychosocial evaluationEmployment history, education, social supportX—Laboratory dataElectrolytes, BUN, creatinine, albumin, hemoglobin, hematocrit, iron, PTH, lipidsXXResidual renal functionTimed urine collection for urea/creatinineXXMedicationsList of medicationsXXVascular access historyType and location of current access, number of prior temporary and permanent accessesX—Vascular access eventsType, location at start and end of interval; all changes in access status; procedures, creation/placement—XHospitalizationsDates, diagnosis, and procedures for each interval hospitalization—XOutpatient eventsDates, diagnosis, and procedures for each outpatient encounter—XInterval statusDialysis status at end of reporting interval, date and cause of departures—XQuality of lifeKidney Disease Quality of Life Instrument (KDQOL) (patient questionnaire)—a Medical Questionnaire (MQ) provides baseline information at time of patient enrollment in DOPPSb Interval Summary (IS) provides follow-up information for each 4-month study interval Open table in a new tab Facility practice patterns are measured by comprehensive questionnaires that are completed by the medical director and the nurse manager (or designee) at each dialysis center. These surveys address a wide range of practice and management issues including dialysis prescription, water quality, dialyzer re-use practices, staffing patterns, nutrition, vascular access, and health maintenance Table 3 and are repeated at yearly intervals. In addition, the summary of patient-specific treatments at the facility level provides valuable information about practice patterns.Table 3Facility level data collection itemsAnemia and iron therapyFacility characteristicsMineral metabolismAntihypertensive therapyFacility staffing practicesNurse and technician practicesContinuing education policies/practiceHealth care maintenancePatient turnoverDialysate processing and compositionHospital and outpatient practicesPhysician practicesDialysis doseImmunizationsPre-ESRD practicesDialysis machinesInformation systemsQuality assurance and improvement practicesDialysis practicesInitiation and discontinuation of dialysisScheduling practicesDialyzer re-useInsurance policiesSocial service practicesDialyzersLaboratory testingVascular accessDietitian and nutrition practicesLocal dialysis marketWater treatment and surveillance Open table in a new tab The same data collection instruments are used in each country, with minor modifications as appropriate (i.e., incorporation of local terminology, deletion of answer choices known to be unavailable). The questionnaires were translated from American English to French, German, Italian, Japanese, Spanish, and Queen's English. In each country, the translated questionnaires were reviewed for meaning and context by nephrologists and pretested in dialysis centers not selected for the study. The primary goal of the study is to find associations between the observed variation in practice patterns and patient-level outcomes, while accounting for patient case-mix and facility clustering effects. A secondary goal is to describe dialysis patients and practices within each country. The study was designed to investigate associations between multiple practice patterns and four specific outcomes (mortality, hospitalization, quality of life, and vascular access events). However, for design purposes, the size of the facility and patient samples was planned based on finding differences in mortality. In Europe, with the smallest sample size per country, the study has a designed power of 90% to detect a mortality difference of 4% between two equal size aggregated groups of dialysis units with different practice patterns of interest. The facility and patient sample from the entire worldwide study will achieve higher power to detect smaller mortality differences. In general, the power to detect clinically important differences in the other major outcomes by differences in treatment practices should be greater than for the dichotomous and rarer outcome of death. A database is maintained at the central coordination and analysis center. A series of range and logical data checks are performed programmatically. Multivariate regression techniques are planned for analyses of the major outcome variables. Explanatory covariates will include both patient variables such as age, sex, race, and comorbid factors, and facility variables such as practice patterns. Facility level factors will be evaluated using techniques that account for facility clustering. Analyses will be conducted to consider both facility-level outcomes (such as census-based mortality rates) and outcomes among only the sampled patients. In this report, selected facility and patient characteristics are expressed using standard descriptive statistics. Statistical comparisons across countries were performed using linear regression for continuous variables and logistic regression for categorical variables. Patient data for the United States were drawn from the randomly selected group of facilities. The United States was arbitrarily selected as the reference group as it contains the largest number of study facilities and patients. A P-value less than 0.05 is considered statistically significant. The study was initiated sequentially in the United States, then Europe, and finally Japan, between 1996 and 1999 and is currently (September 1999) ongoing in all countries. The DOPPS sample is intended to represent the patients (approximately 435,000) and hemodialysis facilities (approximately 8000) in the seven participating countries Table 1. The DOPPS facility sample currently consists of 327 hemodialysis centers. These centers had a total census of 24,392 patients at the start of the study. Thus far, a sample of 10,332 patients have entered detailed longitudinal study (including departed and replacement patients). The overall facility acceptance rate was 77.4%. Figure 1 shows the geographic distribution of the participating dialysis facilities. Table 4 displays basic characteristics of the DOPPS dialysis facilities by country. Large variation in facility size (number of patients treated) was found in all countries, as indicated by the range between the minimum and maximum size. Although facilities treating a small number of patients were not eligible, at least one center fell below the threshold by the time of the initial census. On average, facilities are largest in Japan and the United States and smallest in Spain. Institutionally based facilities (e.g., hospital, university, academic, public) predominate in all countries except Germany, the United States, and Japan where free-standing facilities are more common. Dialysis facility size and type are examples of characteristics (practice patterns) that could plausibly influence patient outcomes.Table 4Characteristics of enrolled DOPPS dialysis facilities by country, 1996–1999Number of hemodialysis patientsType of facilityCountryEnrolled facilitiesMeanMedianMinaSmall units excluded from studyMaxFreestanding (%)bFreestanding includes clinics, private and other non-institutionally based facilitiesInstitutional (%)cInstitutional includes hospital-based, university, academic, and public facilitiesFrance206459321114060Germany206055301108020Italy206554252194060Japan669671264265644Spain20535025905050United Kingdom206258261605050United States1618170153697129a Small units excluded from studyb Freestanding includes clinics, private and other non-institutionally based facilitiesc Institutional includes hospital-based, university, academic, and public facilities Open table in a new tab Table 5 shows basic characteristics of prevalent hemodialysis patients by country. Patient age varies significantly across countries. The average patient age is lowest in the United Kingdom and Japan and highest in Italy and Spain. The majority of hemodialysis patients in all countries are men. However, the percentage of male hemodialysis patients is lower in the United States than all other countries, significantly so for France, Japan, Spain, and the United Kingdom. For example, the odds of a patient being male are 36% higher in Japan than the United States and 39% higher in the United Kingdom than the United States, adjusted for differences in age and diabetes mellitus as the ca