End-of-life and palliative care provision to adults with congenital heart disease: mortality follow-back study using administrative data

Background

Although many adults with congenital heart disease (CHD) still die prematurely, end-of-life care for these patients receives limited attention. There are indications that current care provision at the end of life is burdensome, expensive, and not in line with patients' needs and preferences. We sought to analyse end-of-life care in adult CHD patients to determine whether health services need to be optimized.

Purpose

This study aimed to describe patterns of healthcare consumption of adults with CHD who died in the last year of life.

Methods

This retrospective mortality follow-back study used data of the BELgian COngenital heart disease Database combining Administrative and Clinical data (BELCODAC), including individually linked healthcare claims, death certificates and clinical data from adults with CHD in Flanders (Belgium). For this study, adults with CHD who died between 2007 and 2016 from any cause except sudden death, accident or violence, were selected for inclusion. Accidental, violent, and sudden deaths were identified based on causes of death and healthcare use in the last 3 months of life. Healthcare consumption was based on nomenclature codes derived from healthcare claims data.

Results

A total of 327 eligible patients (median age: 58 y; 54% women; 43% mild CHD; 45% moderate CHD; 11% complex CHD; 49% cardiovascular cause of death) were identified. During the last year of life, healthcare use increased substantially (Fig. 1). During the last month of life, 54% of patients were hospitalised, 55% visited the emergency department, and 15% were admitted to an intensive care unit at least once (Fig. 2). A total of 8% and 5% of patients underwent heart surgery or catherization in the last month of life, respectively. Furthermore, 70% of patients had at least one encounter with a general practitioner and 11% with a CHD specialist in the last month of life. Specialist palliative care was provided to 13% of patients in the last month of life.

When looking at the subgroup of patients with CHD that died due to a cardiovascular cause, proportions of patients that were hospitalised or had visits at the emergency department or intensive care unit in the last month of life were similar (Fig. 2). However, these patients underwent more heart surgeries (11%) and catherizations (8%), had more encounters with CHD specialists (15%), and received remarkably less specialized palliative care (4%) in the last month of life.

Conclusion

Resource utilization increased substantially during the last year of life, resulting in high acute healthcare consumption in the last month of life. It is remarkable that only a minority of patients received palliative care, especially when looking at patients who died due to a cardiovascular cause. Our findings motivate the need to assess if and how end-of-life is planned for adults with CHD. Future studies using qualitative analyses and survey methodology are needed to optimize the management of end-of-life care.

Funding Acknowledgement

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): Research Foundation Flanders, European Society of Cardiology, Koning Boudewijnstichting, National Foundation on Research in Pediatric Cardiology, Swedish Research Council for Health, Working Life and Welfare-FORTE

Healthcare use at the end of life of patients with congenital heart disease: does heart failure matter?

Funding Acknowledgements

Type of funding sources: Foundation. Main funding source(s): Research Foundation Flanders (to PM, EG, and LVB)

European Society of Cardiology (Nursing Training Grant to LVB)

Background

Heart failure (HF) is a common cause of morbidity and mortality in patients with congenital heart disease (CHD). Although limited in scope, previous studies suggest that patients with heart failure follow a specific end-of-life trajectory with episodes of serious complications, which may impact the patterns of care as death approaches.

Aims

The study aims to identify differences in characteristics and patterns of care in the last year of life in deceased CHD patients with and without HF.

Methods

This retrospective study used data of deceased adult patients included in the BELgian COngenital heart disease Database combining Administrative and Clinical data (BELCODAC). To describe patterns of care in the last year of life, we captured information about hospitalisations, emergency department visits, and visits to the general practitioner using nomenclature codes. Heart failure was identified as having HF as cause of death and/or at least one prescription of a loop diuretic in the last year of life. Sensitivity analyses with a stricter definition for HF (HF as cause of death or ≥ 1 prescription of a loop diuretic combined with a prescription of digoxin, dopamine, dobutamine, other non-glycoside stimulants, metoprolol, bisoprolol, carvedilol, aldosterone antagonists, ACE inhibitors or ARBs) were performed as well.

Results

During the period 2007–2016, 390 adults with CHD died, of which 170 patients with HF (44%). Patients with HF were older, died more often due to a cardiovascular cause of death, and had more complex heart lesions, compared to patients without HF (Table 1). While the number of emergency department visits and hospitalisations in the last year was similar, patients with HF had almost twice as much monthly visits at the general practitioner in their last year of life (Table 1). As shown in Figure 1, the mean number of hospitalisations and emergency department visits increased in a similar fashion throughout the last year of life, but the pattern of general practitioner visits was substantially different for patients with and without HF. The sensitivity analyses, in which a stricter definition for HF was used, yield very similar results. In these analyses, the difference in mean monthly hospitalisations was also significant between the two groups.

Conclusions

This study shows clinically important differences in characteristics and patterns of care of deceased patients with CHD with and without heart failure. Patients with HFhave different needs and should receive a tailored approach at the end of life. Future research is needed to understand these differences and investigate these patients' end-of-life care needs in more detail.

Funding acknowledgments: This work was supported by Research Foundation Flanders; European Society of Cardiology; the King Baudouin Foundation; the National Foundation on Research in Pediatric Cardiology; and the Swedish Research Council for Health, Working Life and Welfare-FORTE.

Provision of palliative care to adults with congenital heart disease at the end of life

Funding Acknowledgements

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): This work was supported by Research Foundation Flanders [grant numbers G097516N to PM, 12E9816N to EG and 1154719N to LVB]; the King Baudouin Foundation (Fund Joseph Oscar Waldmann-Berteau & Fund Walckiers Van Dessel); the National Foundation on Research in Pediatric Cardiology; and the Swedish Research Council for Health, Working Life and Welfare -FORTE (grant number STYA-2018/0004).

OnBehalf

BELCODAC consortium

Background

Although recent position papers have discussed and advocated for the integration of palliative care in the treatment course of adults with congenital heart defects (CHD), empirical studies reporting to what extent palliative care is currently provided, are still lacking.

Purpose

(1) To explore the current provision of palliative care to adults with CHD in the last 6 months of their life; and (2) to describe the profile of patients who received palliative care.

Methods

In this retrospective study, data of deceased adult patients included in the BELgian COngenital heart disease Database combining Administrative and Clinical data (BELCODAC) were analysed. Palliative care provision (i.e., admission to palliative care ward, or palliative care at home) was identified using nomenclature codes. The level of anatomical complexity was based on the Bethesda classification. Descriptive analyses were performed.

Results

During the period 2006-2016, 480 adults with CHD died (mean age: 54.4y; 45% simple CHD, 43% moderate CHD, 12% complex CHD). We identified that 75 patients (16%) had at least one nomenclature code linked to palliative care in the last 6 months of their life. More specifically, 16 patients were admitted to an inpatient palliative care service and 67 patients received palliative care at home. Of the patients who received palliative care at home, 40 patients were cared for by a multidisciplinary team specialized in palliative care provision and 59 patients received care from nurses and/or general practitioners while being recognized as a palliative patient. A total of 8 patients received palliative care both at the inpatient palliative care service and at home.

Of the 75 patients receiving palliative care, 44 (59%) had a neoplasm as the primary cause of death and a cardiac cause of death was reported for 10 patients (13%) (see Figure 1). The mean age of patients receiving palliative care was 57.9 years. Most patients receiving palliative care had a simple CHD (n = 40; 53%), 29 patients (39%) had a moderate lesion, and 6 patients (8%) had a complex lesion. That means that, respectively, 19%, 14%, and 11% of all deceased patients with a simple, moderate, and complex heart lesion received palliative care.

Conclusions

This is the first exploratory study on palliative care in adults with CHD. About one in six patients who died received palliative care. Of those who received palliative care, the cause of death was in most cases of a non-cardiac nature. Further research is needed to investigate the care trajectories and care needs of adults with CHD in the last months of life.

Figure 1. Causes of death of adults with CHD who received palliative care in the last 6 months of life (n = 75).

Does kidney transplantation with a standard or expanded criteria donor improve patient survival? Results from a Belgian cohort

BACKGROUND

Changes in recipient and donor factors have reopened the question of survival benefits of kidney transplantation versus dialysis.

METHODS

We analysed survival among 3808 adult Belgian patients waitlisted for a first deceased donor kidney transplant from 2000 to 2012. The primary outcome was mortality during the median waiting time plus 3 years of follow-up after transplantation or with continued dialysis. Outcomes were analysed separately for standard criteria donor (SCD) and expanded criteria donor (ECD) kidney transplants. We adjusted survival analyses for recipient age (20-44, 45-64 and ≥65 years), sex and diabetes as the primary renal disease.

RESULTS

Among patients ≥65 years of age, only SCD transplantation provided a significant survival benefit compared with dialysis, with a mortality of 16.3% [95% confidence interval (CI) 13.2-19.9] with SCD transplantation, 20.5% (95% CI 16.1-24.6) with ECD transplantation and 24.6% (95% CI 19.4-29.5) with continued dialysis. Relative mortality risk was increased in the first months after transplantation compared with dialysis, with equivalent risk levels reached earlier with SCD than ECD transplantation in all age groups.

CONCLUSIONS

The results of this study suggest that older patients might gain a survival benefit with SCD transplantation versus dialysis, but any survival benefit with ECD transplantation versus dialysis may be small.

Cysticercosis and taeniasis cases diagnosed at two referral medical institutions, Belgium, 1990 to 2015

Background

Few case reports on human infections with the beef tapeworm Taenia saginata and the pork tapeworm, Taenia solium, diagnosed in Belgium have been published, yet the grey literature suggests a higher number of cases.

Aim

To identify and describe cases of taeniasis and cysticercosis diagnosed at two Belgian referral medical institutions from 1990 to 2015.

Methods

In this observational study we retrospectively gathered data on taeniasis and cysticercosis cases by screening laboratory, medical record databases as well a uniform hospital discharge dataset.

Results

A total of 221 confirmed taeniasis cases were identified. All cases for whom the causative species could be determined (170/221, 76.9%) were found to be T. saginata infections. Of those with available information, 40.0% were asymptomatic (26/65), 15.4% reported diarrhoea (10/65), 9.2% reported anal discomfort (6/65) and 15.7% acquired the infection in Belgium (11/70). Five definitive and six probable cases of neurocysticercosis (NCC), and two cases of non-central nervous system cysticercosis (non-CNS CC) were identified. Common symptoms and signs in five of the definitive and probable NCC cases were epilepsy, headaches and/or other neurological disorders. Travel information was available for 10 of the 13 NCC and non-CNS CC cases; two were Belgians travelling to and eight were immigrants or visitors travelling from endemic areas.

Conclusions

The current study indicates that a non-negligible number of taeniasis cases visit Belgian medical facilities, and that cysticercosis is occasionally diagnosed in international travellers.

P1630IS THERE ALWAYS A SURVIVAL BENEFIT WITH KIDNEY TRANSPLANTATION? RESULTS FROM A BELGIAN COHORT

Background and Aims

Older studies have shown a survival benefit with kidney transplantation compared to dialysis, even for patients older than 60 years. However, due to important evolutions such as older recipient age and the use of less-than-optimal quality donors, it is unclear if the survival benefit with transplantation still holds true nowadays.

Method

Patient survival was analyzed for 3808 Belgian patients waitlisted for a first deceased donor kidney transplant between 2000 to 2012. Patients were divided into age categories (20-44y, 45-64y, ≥65y). Primary outcome was the comparison of mortality during median waiting time plus 3 years follow-up, either after transplantation or when remaining on dialysis. Outcomes were analyzed separately for those receiving a standard criteria donor (SCD) or an expanded criteria donor (ECD) transplant. The survival analyses were adjusted for age, sex and primary renal disease.

Results

Among patients ≥ 65 years old, only SCD transplantation provided a significant survival benefit compared to dialysis: mortality was 16.3 % (95 % CI: 13.2–19.9 %) with SCD transplantation, 20.5 % (16.1–24.6 %) with ECD transplantation, and 24.6 % (19.5–29.5 %) when remaining on dialysis. Relative mortality risk was increased in the first months after transplantation compared to dialysis, with equal risk levels being reached earlier for SCD than ECD transplantations in all age groups.

Conclusion

This study suggests that older patients have a survival benefit with SCD transplantation versus dialysis, but the survival benefit with ECD transplantation versus dialysis may be small or non-existent.

Counting trees in Random Forests: Predicting symptom severity in psychiatric intake reports

The CEGS N-GRID 2016 Shared Task (Filannino et al., 2017) in Clinical Natural Language Processing introduces the assignment of a severity score to a psychiatric symptom, based on a psychiatric intake report. We present a method that employs the inherent interview-like structure of the report to extract relevant information from the report and generate a representation. The representation consists of a restricted set of psychiatric concepts (and the context they occur in), identified using medical concepts defined in UMLS that are directly related to the psychiatric diagnoses present in the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) ontology. Random Forests provides a generalization of the extracted, case-specific features in our representation. The best variant presented here scored an inverse mean absolute error (MAE) of 80.64%. A concise concept-based representation, paired with identification of concept certainty and scope (family, patient), shows a robust performance on the task.

Selecting relevant features from the electronic health record for clinical code prediction

A multitude of information sources is present in the electronic health record (EHR), each of which can contain clues to automatically assign diagnosis and procedure codes. These sources however show information overlap and quality differences, which complicates the retrieval of these clues. Through feature selection, a denser representation with a consistent quality and less information overlap can be obtained. We introduce and compare coverage-based feature selection methods, based on confidence and information gain. These approaches were evaluated over a range of medical specialties, with seven different medical specialties for ICD-9-CM code prediction (six at the Antwerp University Hospital and one in the MIMIC-III dataset) and two different medical specialties for ICD-10-CM code prediction. Using confidence coverage to integrate all sources in an EHR shows a consistent improvement in F-measure (49.83% for diagnosis codes on average), both compared with the baseline (44.25% for diagnosis codes on average) and with using the best standalone source (44.41% for diagnosis codes on average). Confidence coverage creates a concise patient stay representation independent of a rigid framework such as UMLS, and contains easily interpretable features. Confidence coverage has several advantages to a baseline setup. In our baseline setup, feature selection was limited to a filter removing features with less than five total occurrences in the trainingset. Prediction results improved consistently when using multiple heterogeneous sources to predict clinical codes, while reducing the number of features and the processing time.

A 10-day vacancy period after cleaning and disinfection has no effect on the bacterial load in pig nursery units

BACKGROUND

Biosecurity measures such as cleaning, disinfection and a vacancy period between production cycles on pig farms are essential to prevent disease outbreaks. No studies have tested the effect of a longer vacancy period on bacterial load in nursery units.

METHODS

The present study evaluated the effect of a 10-day vacancy period in pig nursery units on total aerobic flora, Enterococcus spp., Escherichia coli, faecal coliforms and methicillin resistant Staphylococcus aureus (MRSA). Three vacancy periods of 10 days were monitored, each time applied in 3 units. The microbiological load was measured before disinfection and at 1, 4, 7 and 10 days after disinfection.

RESULTS

No significant decrease or increase in E. coli, faecal coliforms, MRSA and Enterococcus spp. was noticed. Total aerobic flora counts were the lowest on day 4 after disinfection (i.e. 4.07 log CFU/625 cm2) (P < 0.05), but the difference with other sampling moments was limited (i.e. 0.6 log CFU/625 cm2) and therefore negligible. Furthermore, this observation on day 4 was not confirmed for the other microbiological parameters. After disinfection, drinking nipples were still mostly contaminated with total aerobic flora (i.e. 5.32 log CFU/625 cm2) and Enterococcus spp. (i.e. 95 % of the samples were positive) (P < 0.01); the feeding troughs were the cleanest location (total aerobic flora: 3.53 log CFU/625 cm2 and Enterococcus spp.: 50 % positive samples) (P < 0.01).

CONCLUSIONS

This study indicates that prolonging the vacancy period in nursery units to 10 days after disinfection with no extra biosecurity measures has no impact on the environmental load of total aerobic flora, E. coli, faecal coliforms, MRSA and Enterococcus spp..

Comparison of competitive exclusion with classical cleaning and disinfection on bacterial load in pig nursery units

Background

Colonisation of the environment of nursery units by pathogenic micro-organisms is an important factor in the persistence and spread of endemic diseases in pigs and zoonotic pathogens. These pathogens are generally controlled by the use of antibiotics and disinfectants. Since an increasing resistance against these measures has been reported in recent years, methods such as competitive exclusion (CE) are promoted as promising alternatives.

Results

This study showed that the infection pressure in CE units after microbial cleaning was not reduced to the same degree as in control units. Despite sufficient administration of probiotic-type spores, the analysed bacteria did not decrease in number after 3 production rounds in CE units, indicating no competitive exclusion. In addition, no differences in feed conversion were found between piglets raised in CE and control units in our study. Also, no differences in faecal consistency (indicator for enteric diseases) was noticed.

Conclusion

These results indicate that the CE protocol is not a valuable alternative for classical C&D.

Data integration of structured and unstructured sources for assigning clinical codes to patient stays

OBJECTIVE

Enormous amounts of healthcare data are becoming increasingly accessible through the large-scale adoption of electronic health records. In this work, structured and unstructured (textual) data are combined to assign clinical diagnostic and procedural codes (specifically ICD-9-CM) to patient stays. We investigate whether integrating these heterogeneous data types improves prediction strength compared to using the data types in isolation.

METHODS

Two separate data integration approaches were evaluated. Early data integration combines features of several sources within a single model, and late data integration learns a separate model per data source and combines these predictions with a meta-learner. This is evaluated on data sources and clinical codes from a broad set of medical specialties.

RESULTS

When compared with the best individual prediction source, late data integration leads to improvements in predictive power (eg, overall F-measure increased from 30.6% to 38.3% for International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic codes), while early data integration is less consistent. The predictive strength strongly differs between medical specialties, both for ICD-9-CM diagnostic and procedural codes.

DISCUSSION

Structured data provides complementary information to unstructured data (and vice versa) for predicting ICD-9-CM codes. This can be captured most effectively by the proposed late data integration approach.

CONCLUSIONS

We demonstrated that models using multiple electronic health record data sources systematically outperform models using data sources in isolation in the task of predicting ICD-9-CM codes over a broad range of medical specialties.

Comparison of sampling procedures and microbiological and non-microbiological parameters to evaluate cleaning and disinfection in broiler houses

Cleaning and disinfection of the broiler stable environment is an essential part of farm hygiene management. Adequate cleaning and disinfection is essential for prevention and control of animal diseases and zoonoses. The goal of this study was to shed light on the dynamics of microbiological and non-microbiological parameters during the successive steps of cleaning and disinfection and to select the most suitable sampling methods and parameters to evaluate cleaning and disinfection in broiler houses. The effectiveness of cleaning and disinfection protocols was measured in six broiler houses on two farms through visual inspection, adenosine triphosphate hygiene monitoring and microbiological analyses. Samples were taken at three time points: 1) before cleaning, 2) after cleaning, and 3) after disinfection. Before cleaning and after disinfection, air samples were taken in addition to agar contact plates and swab samples taken from various sampling points for enumeration of total aerobic flora, Enterococcus spp., and Escherichia coli and the detection of E. coli and Salmonella. After cleaning, air samples, swab samples, and adenosine triphosphate swabs were taken and a visual score was also assigned for each sampling point. The mean total aerobic flora determined by swab samples decreased from 7.7±1.4 to 5.7±1.2 log CFU/625 cm2 after cleaning and to 4.2±1.6 log CFU/625 cm2 after disinfection. Agar contact plates were used as the standard for evaluating cleaning and disinfection, but in this study they were found to be less suitable than swabs for enumeration. In addition to measuring total aerobic flora, Enterococcus spp. seemed to be a better hygiene indicator to evaluate cleaning and disinfection protocols than E. coli. All stables were Salmonella negative, but the detection of its indicator organism E. coli provided additional information for evaluating cleaning and disinfection protocols. Adenosine triphosphate analyses gave additional information about the hygiene level of the different sampling points.

Fine-grained emotion detection in suicide notes: a thresholding approach to multi-label classification

We present a system to automatically identify emotion-carrying sentences in suicide notes and to detect the specific fine-grained emotion conveyed. With this system, we competed in Track 2 of the 2011 Medical NLP Challenge,14 where the task was to distinguish between fifteen emotion labels, from guilt, sorrow, and hopelessness to hopefulness and happiness.

Since a sentence can be annotated with multiple emotions, we designed a thresholding approach that enables assigning multiple labels to a single instance. We rely on the probability estimates returned by an SVM classifier and experimentally set thresholds on these probabilities. Emotion labels are assigned only if their probability exceeds a certain threshold and if the probability of the sentence being emotion-free is low enough. We show the advantages of this thresholding approach by comparing it to a naïve system that assigns only the most probable label to each test sentence, and to a system trained on emotion-carrying sentences only.

Value of transcranial Doppler examination in the diagnosis of cerebral vasospasm after subarachnoid hemorrhage

In 21 patients with subarachnoid hemorrhage secondary to ruptured intracranial aneurysms, we performed serial neurological evaluations, transcranial Doppler examinations, and cerebral blood flow (CBF) determinations. We classified 8 patients as having vasospasm (delayed neurological deterioration, appropriate reduction of CBF) and 13 patients as having no spasm on the basis of this information. Transcranial Doppler flow velocities in the middle cerebral artery and the anterior cerebral artery were significantly elevated for the group with vasospasm on posthemorrhage Days 4 through 12. Elevation of transcranial Doppler velocities preceded clinical signs of cerebral ischemia. The maximal transcranial Doppler flow velocities achieved were compared on the basis of the extent of clot on early computed tomographic (CT) scans. The mean anterior cerebral artery flow velocities were significantly different between CT Grades II and III. The initial transcranial Doppler flow velocities were compared on the basis of the patient's Hunt and Hess grade upon admission. The flow velocities for Grade V patients were significantly lower than those for Grade IV patients. Transcranial Doppler flow velocities were compared with arteriographically observed anterior cerebral artery and middle cerebral artery radii in 12 instances. The correlation was poor, but the data should be interpreted cautiously in view of the small number of arteriograms. We conclude that transcranial Doppler examination has considerable potential in the early diagnosis of delayed ischemic neurological deficit (clinical vasospasm) in patients with subarachnoid hemorrhage.