Identification in Administrative Databases of Women Dying of Breast Cancer

The impossibility of mortality evaluation of skin cancer screening in Germany based on health insurance data: a case-control study

BACKGROUND

The aim of this paper was to perform a mortality evaluation of skin cancer screening (SCS) in Germany using General Local Health Insurance Fund (AOK) data.

METHODS

AOK-insured men and women aged 35-99 years who died of cutaneous malignant melanoma in 2015-2016 were identified. Controls were AOK-insured people who survived to the end of each case's year of death. For each case, 10 controls were matched. The SCS history of each individual was reconstructed using the billing codes 01745 and 01746.

RESULTS

In total, 1037 melanoma deaths and 10,370 controls were included. Cumulative SCS prevalence increased among controls over calendar years, as expected (males and females, 2009: 13.5% and 12.5%; 2015: 52.1% and 55.1%). In contrast, among cases, cumulative SCS prevalence was already high in 2009 and did not show a monotonic increase over the years of diagnosis. Of the 1037 melanoma deaths, 224 (21.6%) had at least one SCS settled in the 12 months after diagnosis.

DISCUSSION

A mortality evaluation with health insurance data alone is not possible because SCS billing codes are not only used for real SCS but also for occasion-related diagnostic work-up of abnormal skin findings. A mortality evaluation with health insurance data requires an individual linking with data of the screening physician and the cancer registries.

Machine-learning model to predict the cause of death using a stacking ensemble method for observational data

OBJECTIVE

Cause of death is used as an important outcome of clinical research; however, access to cause-of-death data is limited. This study aimed to develop and validate a machine-learning model that predicts the cause of death from the patient's last medical checkup.

MATERIALS AND METHODS

To classify the mortality status and each individual cause of death, we used a stacking ensemble method. The prediction outcomes were all-cause mortality, 8 leading causes of death in South Korea, and other causes. The clinical data of study populations were extracted from the national claims (n = 174 747) and electronic health records (n = 729 065) and were used for model development and external validation. Moreover, we imputed the cause of death from the data of 3 US claims databases (n = 994 518, 995 372, and 407 604, respectively). All databases were formatted to the Observational Medical Outcomes Partnership Common Data Model.

RESULTS

The generalized area under the receiver operating characteristic curve (AUROC) of the model predicting the cause of death within 60 days was 0.9511. Moreover, the AUROC of the external validation was 0.8887. Among the causes of death imputed in the Medicare Supplemental database, 11.32% of deaths were due to malignant neoplastic disease.

DISCUSSION

This study showed the potential of machine-learning models as a new alternative to address the lack of access to cause-of-death data. All processes were disclosed to maintain transparency, and the model was easily applicable to other institutions.

CONCLUSION

A machine-learning model with competent performance was developed to predict cause of death.

Implementation of an algorithm for the identification of breast cancer deaths in German health insurance claims data: a validation study based on a record linkage with administrative mortality data

OBJECTIVE

To adapt a Canadian algorithm for the identification of female cases of breast cancer (BC) deaths to German health insurance claims data and to test and validate the algorithm by comparing results with official cause of death (CoD) data on the individual and the population level.

DESIGN

Validation study, secondary data, medical claims.

SETTING

Claims data of two statutory health insurance providers (SHIs) for inpatient and outpatient care, CoD added via record linkage with epidemiological cancer registry (ECR).ParticipantsAll women insured with the two SHIs and who deceased in the period 2006-2013, were residents of North Rhine Westphalia (NRW) and were linked with ECR data: n=22 413.

MAIN OUTCOME MEASURES

Based on inpatient and outpatient diagnoses in the year before death, six algorithms were derived and the accordance of the algorithm-based CoD with the official CoD was evaluated calculating specificity, sensitivity, negative and positive predictive values (NPV, PPV). Furthermore, algorithm-based age-specific BC mortality rates covering several calendar years were calculated for the entire insured female population and compared with official national rates.

RESULTS

Our final algorithm, derived from the NRW subsample, comprised codes indicating the presence of BC, metastases, a terminal illness phase and the absence of codes for other tumours. Overall, specificity, sensitivity, NPV and PPV of this algorithm were 97.4%, 91.3%, 98.9% and 81.7%, respectively. In the age range 40-80 years, sensitivity and PPV slightly decreased with increasing age. Algorithm-based age-specific BC mortality rates agreed well with official rates except for the age group 85 years and older.

CONCLUSIONS

The algorithm-based identification of BC deaths in German claims data is feasible and valid, except for higher ages. The algorithm to ascertain BC mortality rates in an epidemiological study seems applicable when information on the official CoD is not available in the original database.

The economic burden of end-of-life care in metastatic breast cancer

OBJECTIVE

To assess end-of-life (EOL) total healthcare costs and resource utilization during the last 6 months of claims follow-up among patients with metastatic breast cancer (MBC) who received systemic anti-neoplastic therapy.

METHODS

Newly diagnosed females with MBC initiating treatment January 1, 2003-June 30, 2011 were identified in a large commercial claims database. Two cohorts were defined based on a proxy measure for EOL 1 month prior to the end of last recorded follow-up within the study period: patients who were assumed dead at end of claims follow-up (EOL cohort) and patients who were alive (no-end-of-life [NEOL] cohort). Proxy measures for EOL were obtained from published literature and clinical expert opinion. Cost and resource utilization were evaluated for the 6 months prior to end of claims follow-up. Baseline variables, resource utilization, and costs were compared between cohorts with univariate statistical tests. Adjusted relative risks were calculated for resource utilization measures. A covariate-adjusted generalized linear model evaluated 6-month total healthcare costs.

RESULTS

Of the 3,878 females included, 18.5% (n = 718) met the criteria for EOL. Mean observational time (MBC onset to end of claims follow-up) was shorter for the EOL cohort (EOL, 32 months vs NEOL, 35 months; p < 0.001). In adjusted analyses, the EOL cohort had 4.15 times higher 6-month total healthcare costs (EOL, $72,112 vs NEOL, $17,137; p < 0.001). NEOL month-to-month mean total healthcare costs fluctuated between $2336-$3145, while EOL costs increased steadily from $8,956 in the sixth month prior to death to $19,326 in the last month of life. The adjusted relative risk of inpatient, hospice and emergency department utilization was >2 times higher in the EOL cohort (p < 0.001).

CONCLUSIONS

Potential EOL presented a greater economic burden in the 6 months prior to death. EOL month-to-month costs increased precipitously in the last 2 months of life and were driven by acute inpatient care.

Receipt of Life-Sustaining Treatments for Taiwanese Pediatric Patients Who Died of Cancer in 2001 to 2010: A Retrospective Cohort Study

Aggressive life-sustaining treatments have the potential to be continued beyond benefit, but have seldom been systematically/nationally explored in pediatric cancer patients. Furthermore, factors predisposing children dying of cancer to receive life-sustaining treatments at end of life (EOL) have never been investigated in a population-based study. This population-based study explored determinants of receiving life-sustaining treatments in pediatric cancer patients' last month of life. For this retrospective cohort study, we used administrative data on 1603 Taiwanese pediatric cancer patients who died in 2001 to 2010. Individual patient-level data were linked with encrypted identification numbers from the National Register of Deaths Database, Cancer Registration System database, National Health Insurance claims datasets, and Database of Medical Care Institutions Status. Life-sustaining treatments included intensive care unit (ICU) care, cardiopulmonary resuscitation (CPR), and mechanical ventilation. Associations of patient, physician, hospital, and regional factors with receiving ICU care, CPR, and mechanical ventilation in the last month of life were evaluated by multilevel generalized linear mixed models. In their last month of life, 22.89%, 46.48%, and 61.45% of pediatric cancer patients received CPR, mechanical ventilation, and ICU care, respectively, with no significant decreasing trends from 2001 to 2010. Patients were more likely to receive all three identified life-sustaining treatments at EOL if they were diagnosed with a hematologic malignancy or a localized disease, died within 1 year of diagnosis, and received care from a pediatrician. Receipt of ICU care or mechanical ventilation increased with increasing EOL-care intensity of patients' primary hospital, whereas use of mechanical ventilation decreased with increasing quartile of hospice beds in the patients' primary hospital region. Taiwanese pediatric cancer patients received aggressive life-sustaining treatments in the month before death. Healthcare policies and interventions should aim to help pediatricians treating at-risk pediatric cancer patients and hospitals with a tendency to provide aggressive EOL treatments to avoid the expense of life-sustaining treatments when chance of recovery is remote and to devote resources to care that produces the greatest benefits for children, parents, and society.

The Association Between Home Palliative Care Services and Quality of End-of-Life Care Indicators in the Province of Québec

CONTEXT

In Canada, governments have increased spending on home care to promote better end-of-life care. In the province of Québec, Canada, home palliative care (PC) services (HPCS) are provided by Public Local Community-Based Health Care Service providers (Centres Locaux de Services Communautaires [CLSC]) with universal coverage. Accordingly, there should be no regional variations of these services and their effect on quality of end-of-life PC (QEoLPC) indicators.

OBJECTIVES

To test if all the CLSCs provided the same level of HPCS to cancer patients in the province of Québec, Canada, and the association between level of HPCS and QEoLPC indicators.

METHODS

Characteristics of 52,316 decedents with cancer were extracted from administrative databases between 2003 and 2006. Two gender-specific "adjusted performance of CLSCs in delivering HPCS" models were created using gender-specific hierarchical regression adjusted for patient and CLSC neighborhood characteristics. Using the same approach, the strength of the association between the adjusted performance of CLSCs in delivering HPCS and the QEoLPC indicators was estimated.

RESULTS

Overall, 27,255 (52.1%) decedents had at least one HPCS. Significant variations in the adjusted performance of CLSC in delivering HPCS were found. Higher performance led to a lower proportion of men having more than one emergency room visit during the last month of life (risk ratio [RR] 0.924; 95% CI 0.867-0.985), and for women, a higher proportion dying at home (RR 2.255; 95% CI 1.703-2.984) and spending less time in hospital (RR 0.765; 95% CI 0.692-0.845).

CONCLUSION

Provision of HPCS remained limited in Québec, but when present, they were associated with improved QEoLPC indicators.

Measuring intensity of end of life care: a systematic review

Background

Many studies have measured the intensity of end of life care. However, no summary of the measures used in the field is currently available.

Objectives

To summarise features, characteristics of use and reported validity of measures used for evaluating intensity of end of life care.

Methods

This was a systematic review according to PRISMA guidelines. We performed a comprehensive literature search in Ovid Medline, Embase, The Cochrane Library of Systematic Reviews and reference lists published between 1990-2014. Two reviewers independently screened titles, abstracts, full texts and extracted data. Studies were eligible if they used a measure of end of life care intensity, defined as all quantifiable measures describing the type and intensity of medical care administered during the last year of life.

Results

A total of 58 of 1590 potentially eligible studies met our inclusion criteria and were included. The most commonly reported measures were hospitalizations (n = 44), intensive care unit admissions (n = 39) and chemotherapy use (n = 27). Studies measured intensity of care in different timeframes ranging from 48 hours to 12 months. The majority of studies were conducted in cancer patients (n = 31). Only 4 studies included information on validation of the measures used. None evaluated construct validity, while 3 studies considered criterion and 1 study reported both content and criterion validity.

Conclusions

This review provides a synthesis to aid in choosing intensity of end of life care measures based on their previous use but simultaneously highlights the crucial need for more validation studies and consensus in the field.

Patterns of community-based opioid prescriptions in people dying of cancer

CONTEXT

Studies of opioid use in cancer patients have been cross-sectional or have focused on mean consumption over a specific time interval.

OBJECTIVES

This study aimed to determine the temporal pattern of prescribed opioids at a population level.

METHODS

Using Quebec administrative databases, we ascertained details of cancer-related deaths and filled community-based opioid prescriptions (COPs) in 48,420 decedents from 2003 to 2006.

RESULTS

Using group-based trajectory modeling, based on when people started to fill COPs, our population-based study demonstrated patterns of filled COPs with six distinct trajectories. An earlier start in opioid consumption resulted in a higher group average morphine daily dose; those who were already filling COPs at study entry (5.2%) had a final dose of more than 300mg by the time of death. Remarkably, 58.8% of people had not filled COPs with a biweekly average greater than 1mg earlier than two weeks before death, marking the end of follow-up. Breast cancer in women, prostate or colorectal cancer in men, and younger age and multiple myeloma in both sexes were positively associated with earlier filling of COPs.

CONCLUSION

Patients dying of cancer require increasing doses of opioids over time; although we cannot distinguish the relative contributions of disease progression and opioid tolerance, age and certain cancers seem related to this phenomenon. Given the potentially prohibitive cost of prospective epidemiological studies, more elaborate clinical administrative databases that include regular pain assessment are necessary to determine optimal opioid use and factors associated with dose increases over time at a population level.

Limited validity of diagnosis codes in Medicare claims for identifying cancer metastases and inferring stage

PURPOSE

Researchers are using diagnosis codes from health claims to identify metastatic disease in cancer patients. The validity of this approach has not been established.

METHODS

We used the linked 2005-2007 Surveillance, Epidemiology and End Results (SEER)-Medicare data to assess the validity of metastasis codes at diagnosis from claims compared with stage reported by SEER cancer registries. The cohort included 80,052 incident breast, lung, and colorectal cancer patients aged 65 years and older. Using gold-standard SEER data, we evaluated sensitivity, specificity, positive predictive value, and negative predictive value of claims-based stage, survival by stage classification, and patient factors associated with stage misclassification using multivariable regression.

RESULTS

For patients with a registry report of distant metastatic cancer, the sensitivity, specificity, and positive predictive value of claims never simultaneously exceeded 80% for any cancer: lung (42.7%, 94.8%, and 88.1%), breast (51.0%, 98.3%, and 65.8%), and colorectal (72.8%, 93.8%, and 68.5%). Misclassification of stage from Medicare claims was significantly associated with inaccurate estimates of stage-specific survival (P < .001). In adjusted analysis, patients who were older, black, or living in low-income areas were more likely to have their stage misclassified in claims.

CONCLUSIONS

Diagnosis codes in Medicare claims have limited validity for inferring cancer stage and metastatic disease.

Delivery of health care at the end of life in cancer patients of four swiss cantons: a retrospective database study (SAKK 89/09)

BACKGROUND

The use of cancer related therapy in cancer patients at the end-of-life has increased over time in many countries. Given a lack of published Swiss data, the objective of this study was to describe delivery of health care during the last month before death of cancer patients.

METHODS

Claims data were used to assess health care utilization of cancer patients (identified by cancer registry data of four participating cantons), deceased between 2006-2008. Primary endpoints were hospitalization rate and delivery of cancer related therapies during the last 30 days before death. Multivariate logistic regression assessed the explanatory value of patient and geographic characteristics.

RESULTS

3809 identified cancer patients were included. Hospitalization rate (mean 68.5%, 95% CI 67.0-69.9) and percentage of patients receiving anti-cancer drug therapies (ACDT, mean 14.5%, 95% CI 13.4-15.6) and radiotherapy (mean 7.7%, 95% CI 6.7-8.4) decreased with age. Canton of residence and insurance type status most significantly influenced the odds for hospitalization or receiving ACDT.

CONCLUSIONS

The intensity of cancer specific care showed substantial variation by age, cancer type, place of residence and insurance type status. This may be partially driven by cultural differences within Switzerland and the cantonal organization of the Swiss health care system.

Patterns of care at end of life for people with primary intracranial tumors: lessons learned

To determine the variability in processes of care in the last 6 months of life experienced by patients dying of primary intracranial tumors and potential predictors of place of death, a death-backwards cohort was assembled using historical data and 1,623 decedents were identified. 90 % of people had ≥ 1 admission to an acute care hospital and 23 % spent ≥ 3 months of their last 6 months of life in acute care. 44 % had ≥ 1 ER visits and 30 % were admitted ≥ 1 times to ICU. Only 18 % had a home visit by a physician. 10 % died at home but 49 % died in hospital, while 40 % died in a palliative care facility. Age, comorbidities, and being diagnosed with grade 4 astrocytoma were associated with greater burden of care. Level of care burden and age were associated with higher odds of dying in a treatment intensive place of death, being diagnosed with grade 4 astrocytoma had opposite effect. Despite valuable research efforts to improve the treatment of primary intracranial tumors that focus on biology, refinements to surgery, radiation, and chemotherapy, there is also room to improve aspects of care at the end of life situation. An integrative approach for this patients' population, from diagnosis to death, could potentially reduce the care burden in the final period on the health care system, patient's family and improve access to a better place of death.

Impact of age on end-of-life care for adult Taiwanese cancer decedents, 2001-2006

BACKGROUND

With increasing patient age in Western countries, evidence indicates a pervasive pattern of decreasing healthcare expenditures and less aggressive medical care, including end-of-life (EOL) care. However, the impact of age on EOL care for Asian cancer patients has not been investigated.

PURPOSE

To explore how healthcare use at EOL varies by age among adult Taiwanese cancer patients.

METHODS

Retrospective cohort study using administrative data among 203,743 Taiwanese cancer decedents, 2001-2006. Age was categorized as 18-64, 65-74, 75-84, and ≥85 years.

RESULTS

Elderly (≥65 years) Taiwanese cancer patients were significantly less likely than those 18-64 years to receive aggressive treatment in their last month of life, including chemotherapy, >1 emergency room visits, >1 hospital admissions, >14 days of hospitalization, hospital death, intensive care unit admission, cardiopulmonary resuscitation, intubation, and mechanical ventilation. However, they were significantly more likely to receive hospice care in their last year of life.

CONCLUSION

Elderly Taiwanese cancer patients at EOL received less chemotherapy, less aggressive management of health crises associated with the dying process, and fewer life-extending treatments, but they were more likely to receive hospice care in their last year and to achieve the culturally highly valued goal of dying at home.

The effect on survival of continuing chemotherapy to near death

BACKGROUND

Overuse of anti-cancer therapy is an important quality-of-care issue. An aggressive approach to treatment can have negative effects on quality of life and cost, but its effect on survival is not well-defined.

METHODS

Using the Surveillance, Epidemiology, and End Results-Medicare database, we identified 7,879 Medicare-enrolled patients aged 65 or older who died after having survived at least 3 months after diagnosis of advanced non-small cell lung cancer (NSCLC) between 1991 and 1999. We used Cox proportional hazards regression analysis, propensity scores, and instrumental variable analysis (IVA) to compare survival among patients who never received chemotherapy (n = 4,345), those who received standard chemotherapy but not within two weeks prior to death (n = 3,235), and those who were still receiving chemotherapy within 14 days of death (n = 299). Geographic variation in the application of chemotherapy was used as the instrument for IVA.

RESULTS

Receipt of chemotherapy was associated with a 2-month improvement in overall survival. However, based on three different statistical approaches, no additional survival benefit was evident from continuing chemotherapy within 14 days of death. Moreover, patients receiving chemotherapy near the end of life were much less likely to enter hospice (81% versus 51% with no chemotherapy and 52% with standard chemotherapy, P < 0.001), or were more likely to be admitted within only 3 days of death.

CONCLUSIONS

Continuing chemotherapy for advanced NSCLC until very near death is associated with a decreased likelihood of receiving hospice care but not prolonged survival. Oncologists should strive to discontinue chemotherapy as death approaches and encourage patients to enroll in hospice for better end-of-life palliative care.

Why do patients with cancer visit the emergency department near the end of life?

BACKGROUND

For patients dying of cancer, a visit to the emergency department can be disruptive, distressing and exhausting. Such visits made near the end of life are considered an indicator of poor-quality cancer care. We describe the most common reasons for visits made to the emergency department during the final six months of life and the final two weeks of life by patients dying of cancer.

METHODS

We performed a descriptive, retrospective cohort study using linked administrative sources of health care data.

RESULTS

Between 2002 and 2005 in Ontario, 91,561 patients died of cancer. Of these, 76,759 patients made 194,017 visits to the emergency department during the final six months of life. Further, 31,076 patients made 36,600 visits to the emergency department during the final two weeks of life. In both periods, the most common reasons were abdominal pain, lung cancer, dyspnea, pneumonia, malaise and fatigue, and pleural effusion.

INTERPRETATION

Many visits made to the emergency department by patients with cancer near the end of life may be avoidable. An understanding of the reasons for such visits could be useful in the development of dedicated interventions for preventing or avoiding their occurrence.

Determinants of aggressive end-of-life care for Taiwanese cancer decedents, 2001 to 2006

PURPOSE

To assess the association between aggressiveness of end-of-life (EOL) care and patient demographics, disease characteristics, primary physician's specialty, hospital characteristics, and availability of health care resources at the hospital and regional levels in Taiwan for a cohort of 210,976 cancer decedents in 2001 to 2006.

METHODS

This retrospective cohort study examined administrative data. Aggressiveness of EOL care was examined by a composite measure adapted from Earle et al. Scores range from 0 to 6, with higher scores indicating more aggressive EOL care.

RESULTS

The mean composite score for aggressiveness of EOL care was 2.04 (mean) +/- 1.26 (standard deviation), increasing from 1.96 +/- 1.26 in 2001 to 2.10 +/- 1.26 in 2006. Each successive year of death significantly increased the composite score. Cancer decedents received more aggressive EOL care if they were male, younger, single, had a higher level of comorbidity, had more malignant and extensive diseases or hematologic malignancies, were cared for by oncologists, and received care in a hospital with a greater density of beds.

CONCLUSION

Controlling for patient demographics and cormorbidity burden, EOL care in Taiwan was more aggressive for patients with cancer with highly malignant and extensive diseases, for patients with oncologists as primary care providers, or in hospitals with abundant health care resources. Health policies should aim to ensure that all patients receive treatments that best meet their individual needs and interests and that resources are devoted to care that produces the greatest health benefits.

Patterns of access to community palliative care services: a literature review

Policies state that access to palliative care should be provided according to principles of equity. Such principles would include the absence of disparities in access to health care that are systematically associated with social advantage. A review of the literature a decade ago identified that patients with different characteristics used community palliative care services in variable ways that appeared inequitable. The objective of this literature review was to review recent literature to identify whether such variability remains. Searching included the use of electronic databases, scrutinizing bibliographies, and hand searching journals. Articles were included if they were published after 1997 (the date of the previous review) up to the beginning of 2008, and if they reported any data that investigated the characteristics of adult patients in relation to their relative utilization of community palliative care services, with reference to a comparator population. Forty-eight studies met the inclusion criteria. Patients still access community palliative care services in variable ways. Those who are older, male, from ethnic minority populations, not married, without a home carer, are socioeconomically disadvantaged, and who do not have cancer are all less likely to access community palliative care services. These studies do not identify the reasons for such variable access, or whether such variability is warranted with reference to clinical need or other factors. Studies tend to focus on access to specialist palliative care services without looking at the complexities of service use. Studies need to move beyond description of utilization patterns, and examine whether such patterns are inequitable, and what is happening in the referral or other processes that may result in such patterns.

Trends in quality of end-of-life care for Taiwanese cancer patients who died in 2000-2006

BACKGROUND

Quality of end-of-life care received by cancer patients has never been explored in an entire Asian country for all ages and cancer groups.

PATIENTS AND METHODS

Retrospective cohort study to examine trends in quality of end-of-life care among a cohort of 242 530 Taiwanese cancer patients who died in 2000-2006.

RESULTS

In the last month of life, cancer care tended to become increasingly aggressive as shown by (i) intensive use of chemotherapy (15.45%-17.28%), (ii) frequent emergency room visits (15.69%-20.99%) and >14-day hospital stays (41.48%-46.20%), (iii) admissions to intensive care units (10.04%-12.41%), and (iv) hospital deaths (59.11%-65.40%). Use of cardiopulmonary resuscitation (13.09%-8.41%), intubation (26.01%-21.07%), and mechanical ventilation (27.46%-27.05%) decreased, whereas use of hospice services increased considerably (7.34%-16.83%). Among those receiving hospice services, rates of referrals to hospice services in the last 3 days of life decreased from 17.88% to 17.13% but remained steady after adjusting for selected covariates.

CONCLUSIONS

The quality of end-of-life care for Taiwanese cancer decedents was substantially inferior to that previously reported and to that recommended as benchmarks for not providing overly aggressive care near the end of life.

Nationwide monitoring of end-of-life care via the Sentinel Network of General Practitioners in Belgium: the research protocol of the SENTI-MELC study

Background

End-of-life care has become an issue of great clinical and public health concern. From analyses of official death certificates, we have societal knowledge on how many people die, at what age, where and from what causes. However, we know little about how people are dying. There is a lack of population-based and nationwide data that evaluate and monitor the circumstances of death and the care received in the final months of life. The present study was designed to describe the places of end-of-life care and care transitions, the caregivers involved in patient care and the actual treatments and care provided to dying patients in Belgium. The patient, residence and healthcare characteristics associated with these aspects of end-of-life care provision will also be studied. In this report, the protocol of the study is outlined.

Methods/Design

We designed a nationwide mortality follow-back study with data collection in 2005 and 2006, via the nationwide Belgian Sentinel Network of General Practitioners (GPs) i.e. an existing epidemiological surveillance system representative of all GPs in Belgium, covering 1.75% of the total Belgian population. All GPs were asked to report weekly, on a standardized registration form, every patient (>1 year) in their practice who had died, and to identify patients who had died "non-suddenly." The last three months of these patients' lives were surveyed retrospectively. Several quality control measures were used to ensure data of high scientific quality.

Discussion

In 2005 and 2006, respectively 1385 and 1305 deaths were identified of which 66% and 63% died non-suddenly. The first results are expected in 2007. Via this study, we will build a descriptive epidemiological database on end-of-life care provision in Belgium, which might serve as baseline measurement to monitor end-of-life care over time. The study will inform medical practice as well as healthcare authorities in setting up an end-of-life care policy. We publish the protocol here to inform others, in particular countries with analogue GP surveillance networks, on the possibilities of performing end-of-life care research. A preliminary analysis of the possible strengths, weaknesses and opportunities of our research is outlined.

Towards using administrative databases to measure population-based indicators of quality of end-of-life care: testing the methodology

This study is concerned with methods to measure population-based indicators of quality end-of-life care. Using a retrospective cohort approach, we assessed the feasibility, validity and reliability of using administrative databases to measure quality indicators of end-of-life care in two Canadian provinces. The study sample consisted of all females who died of breast cancer between 1 January 1998 and 31 December 2002, in Nova Scotia or Ontario, Canada. From an initial list of 19 quality indicators selected from the literature, seven were determined to be fully measurable in both provinces. An additional seven indicators in one province and three in the other province were partially measurable. Tests comparing administrative and chart data show a high level of agreement with inter-rater reliability, confirming consistency in the chart abstraction process. Using administrative data is an efficient, population-based method to monitor quality of care which can compliment other methods, such as qualitative and purposefully collected clinical data.