PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • The study protocol was approved by the Winnipeg Regional Health Authority (WRHA) Research Review Board and the Health Research Ethics Board of the University of Manitoba. We assembled a population-based retrospective cohort of kidney failure and non-kidney failure patients by linking the Manitoba Renal Program (MRP) provincial database, a registry of all Manitoba kidney failure patients, and the Emergency Admission, Discharge, and Transfer (ADT) Database, an administrative registry of all patients presenting to any ED in the Winnipeg Regional Health Authority (WRHA). Linkage was achieved via utilization of unique Personal Health Identifier Number (PHIN). The study population consisted of the entire population within the WRHA, from the period of Jan 1, 2000 to Dec 31, 2009. The WRHA is responsible for the health care of over 700,000 Winnipeg residents and the tertiary care of nearly 500,000 additional residents of Manitoba [9], representing the vast majority of the province’s population. The MRP registry has tracked both incident and prevalent dialysis patients in Manitoba since 1996. The registry excludes patients with acute kidney injury who required temporary dialysis. For the purposes of the present study, we linked data from Jan 1, 2000 to Dec 31, 2009. The registry is complete for all prevalent kidney failure patients in Manitoba and provides the basis for billing and reporting of provincial dialysis information to the Canadian Organ Replacement Registry (CORR), the body responsible for the collection of medical data of patients with kidney failure in Canada [10]. Data elements captured included the patient’s PHIN, date of initial dialysis treatment, and modality. All changes to vital status and dialysis modality are updated weekly at interprofessional rounds and are reported in the MRP and CORR registries [11]. The ADT system has tracked all ED visits in the WRHA from its inception in 1999 until Dec 31, 2009, with the purpose of providing real-time data entry and patient registration uniformly across the entire region. We linked data from Jan 1, 2000 to Dec 31, 2009. The database is known to have captured 100% of all ED presentations in the WRHA over this period. Data elements recorded include the patient’s PHIN, general demographics, address, time and date of presentation, hospital of presentation, mode of arrival, and Canadian Triage and Acuity Scale (CTAS) score [12]. Data are entered in real time by skilled personnel upon patient presentation to the ED triage nurse. Data was linked across databases at the individual level via the patient’s PHIN and patients with non-valid or corrupt PHINs were excluded. All linkages were done on a secure server. Prior to exporting the file for analysis, each patient was given a unique study number and the PHINs were purged from the data, ensuring anonymity of the exported file. The resulting linked data file described ED visits in over 1.2 million unique individuals over the study period. Dialysis status was assessed separately for each ED visit for each patient using the MRP registry. Patients not appearing in the registry were classified as non-dialysis patients for all ED visits. Patients were classified as a dialysis patient if the date of the ED visit for that patient occurred after the date of initial dialysis, as recorded in the MRP registry. Patients who had ED visits before and after initiation of dialysis were classified as non-dialysis patients for ED visits occurring before and as dialysis patients for visits occurring after dialysis start. The vast majority of patients on chronic dialysis in Manitoba are treated with in-centre conventional hemodialysis (CHD, 75%) or peritoneal dialysis (PD, 20%), with a small number undergoing home hemodialysis (HHD, 5%). These proportions are similar to other Canadian provinces. For the purposes of the present analyses, we combined CHD and HHD into one group (HD). We classified kidney transplant patients as non-dialysis patients. Over time, patients undergo modality switches such as changing from transplant to dialysis, or HD to PD (and vice versa). Because the MRP registry records the start and end dates of each modality transition, we attributed each ED visit to the concurrent treatment modality. Demographic data were taken from the ADT system. As this data is also in the MRP registry for patients with kidney failure, the data were compared and reconciled in the small number of cases where disagreement existed. Patient age was calculated for each ED visit based on date of birth and date of ED admission. Normally distributed continuous variables of interest were summarized as mean (SD). Dichotomous variables and outcomes were summarized as percentages. T tests and ANOVA were used to compare normally distributed measures. The Mann-Whitney U test and the Kruskall-Wallis test were used for non-Gaussian distributions. The χ2 test was used to compare dichotomous variables. Secular trends in rates of ED visits: For each year of follow-up, a crude yearly ED visit rate for patients on chronic dialysis was calculated as the number of ED visits by dialysis patients in that year divided by the number of dialysis patients registered in the MRP database at the end of that year. Patients who were on dialysis for less than 6 months of the year in question were classified as non-dialysis patients for that year. Because patients on chronic dialysis were older and had a greater proportion of men than the general Manitoba population, we adjusted the crude rates using direct age and gender standardization referenced to the 2006 adult population of Manitoba. Yearly ED visit rates for non-dialysis patients were calculated as the number of ED visits by non-dialysis patients divided by the estimated adult population of Manitoba in that year. The adult population in Manitoba was estimated for each year using linear interpolation of population data from the 2001, 2006, and 2011 censuses. Clustering of ED visits in relation to day of the week: HD patients typically dialyze three times a week on a Monday/Wednesday/Friday or a Tuesday/Thursday/Saturday schedule, and thus experience two short and one long interdialytic interval (i.e. the Friday to Monday interval or Saturday to Tuesday interval, respectively). To address the question of whether ED visits might be clustered on days following the long interdialytic intervals in HD patients, we calculated the rate of ED visits per patient year occurring on each day of the week, stratified by HD, PD, and non-dialysis status. Rates were compared across days of the week and across strata using Poisson regression. The hypothesis that the day of the week effect was more pronounced in HD than in other categories was tested using a formal interaction term (weekday x kidney failure status). Clustering of ED visits in relation to dialysis start date: To address the question of whether ED visits clustered shortly before or shortly after the initiation of chronic dialytic therapy, we calculated the timing in weeks of each ED visit in relation to the dialysis initiation date for each patient (e.g. visit occurred 21–28 days before; 21–28 days after initiation of dialysis). We then calculated the total number of ED visits for each week preceding and each week following initiation of dialysis. We investigated ED utilization up to 52 weeks prior to the initiation of dialysis and 52 weeks after the initiation of dialysis in both HD and PD patients. The weekly frequencies were expressed as rate per patient-year of follow-up.
rdf:type