Abstract P454: Do Not Resuscitate Orders in Intracerebral Hemorrhage Patients. Impact on Mortality

Stroke ◽  
2021 ◽  
Vol 52 (Suppl_1) ◽  
Author(s):  
Mohammad Rauf A Chaudhry ◽  
Hussan Gill ◽  
Saqib Chaudhry ◽  
Baljinder Singh ◽  
Harathi Bandaru ◽  
...  
Keyword(s):  
Intracerebral Hemorrhage ◽  
The Self ◽  
Racial Groups ◽  
Do Not Resuscitate ◽  
Code Status ◽  
International Classification Of Disease ◽  
Discharge Outcomes ◽  
Racial Ethnic ◽  
Dnr Orders

Introduction/background: Do not resuscitate (DNR) orders have been associated with higher mortality in hospitalized patients which the question if they these patients are victims of the self-fulfilling prophecy; that the odds of their survival is made worse by withholding aggressive treatment. In addition, previous reports show that racial and ethnic minorities tend to opt for more aggressive and lifesaving procedures as compared to Whites. Methods: We analyzed the data from Nationwide Inpatient Sample (2005-2014) for all intracerebral hemorrhage (ICH) patients. DNR code status was identified using the International Classification of Disease, 9th Revision, Clinical Modification codes. Baseline characteristics, discharge outcomes (mortality, length of stay) were compared between the two groups. DNR code status was compared between different racial groups. Results: Of the 884379 patients with ICH, 81968 (9.26%%) had DNR order. ICH patients with DNR order had higher proportion of females (55.1% versus 49.1%, P <.0001) and were older (74.2 years versus 66 years, P <.0001) compared to ICH patients without DNR. The in-hospital mortality was also higher (53.4% versus 23.3%, p≤.0001) among patients with DNR both univariate and multivariate analysis (OR = 3.24 (3.07 -3.41), p<.0001) after adjusting for potential confounders. Whites have a higher rate (11.5% versus 8.08%) of DNR order as compared to other racial/ethnic groups Conclusions: While there may be other explanations at play, the higher mortality and shorter LOS suggest that early DNR orders do lead to the self-fulfilling prophecy. The lower proportion of DNR orders among minorities suggest a sociocultural aspect in accepting the concept of DNR. These two facts raise concerns about what the real vs perceived meaning of DNR orders.

Abstract P399: Do Not Resuscitate Orders in Subarachnoid Hemorrhage Patients. Impact on Mortality

Stroke ◽  
2021 ◽  
Vol 52 (Suppl_1) ◽  
Author(s):  
Mohammad Rauf Chaudhry ◽  
Hussan Gill ◽  
Saqib Chaudhry ◽  
Baljinder Singh ◽  
Harathi Bandaru ◽  
...  
Keyword(s):  
Subarachnoid Hemorrhage ◽  
The Self ◽  
Racial Groups ◽  
Do Not Resuscitate ◽  
Code Status ◽  
International Classification Of Disease ◽  
Discharge Outcomes ◽  
Racial Ethnic ◽  
Dnr Orders

Introduction/background: Do not resuscitate (DNR) orders have been associated with higher mortality in hospitalized patients which the question if they these patients are victims of the self-fulfilling prophecy; that the odds of their survival is made worse by withholding aggressive treatment. In addition, previous reports show that racial and ethnic minorities tend to opt for more aggressive and lifesaving procedures as compared to Whites. Methods: We analyzed the data from Nationwide Inpatient Sample (2005-2014) for all subarachnoid hemorrhage (ICH) patients. DNR code status was identified using the International Classification of Disease, 9th Revision, Clinical Modification codes. Baseline characteristics, discharge outcomes (mortality, length of stay) were compared between the two groups. DNR code status was compared between different racial groups. Results: Of the 325923 patients with SAH, 20127 (6.17%%) had DNR order. SAH patients with DNR order had higher proportion of females (61.7% versus 59.4%, P=0.0048) and were older (70.9 years versus 58.4 years, P <.0001) compared to SAH patients without DNR. The in-hospital mortality was also higher (66.4% versus 22.4%, p≤.0001) among patients with DNR both univariate and multivariate analysis (OR = 5.05 (4.58 -5.56), p<.0001) after adjusting for potential confounders. Whites have a higher rate (7.59% versus 5.84%, P <.0001) of DNR order as compared to other racial/ethnic groups Conclusions: While there may be other explanations at play, the higher mortality and shorter LOS suggest that early DNR orders do lead to the self-fulfilling prophecy. The lower proportion of DNR orders among minorities suggest a sociocultural aspect in accepting the concept of DNR. These two facts raise concerns about what the real vs perceived meaning of DNR orders.

Abstract P456: Myocardial Infarction in Patients With Intracerebral Hemorrhage. Prevalence and Impact on Outcome

Stroke ◽  
2021 ◽  
Vol 52 (Suppl_1) ◽  
Author(s):  
Mohammad Rauf Chaudhry ◽  
Hussan Gill ◽  
Saqib Chaudhry ◽  
Baljinder Singh ◽  
Harathi Bandaru ◽  
...  
Keyword(s):  
Myocardial Infarction ◽  
Intracerebral Hemorrhage ◽  
Nationwide Inpatient Sample ◽  
Hospital Charges ◽  
Discharge Disposition ◽  
International Classification Of Disease ◽  
Myocardial Infraction ◽  
Discharge Outcomes ◽  
Baseline Characteristics

Introduction/background: Comorbidities can potentially affect outcome of patients with intracerebral hemorrhage (ICH). It is unclear what the prevalence of acute myocardial infarction (AMI) and its impact on outcome are in patients with intracerebral hemorrhage. Methods: We analyzed the data from Nationwide Inpatient Sample (2005-2014) for all intracerebral hemorrhage (ICH) patients. AMI was identified using the International Classification of Disease, 9th Revision, Clinical Modification codes. Baseline characteristics, discharge outcomes (mortality, discharge disposition, length of stay and in-hospital charges) were compared between the two groups. Results: Of the 884379 patients with ICH, 27692 (3.13%) had in-hospital myocardial infraction. ICH patients with AMI order had lower proportion of females (47.8% versus 49.7%, P= 0.0028) and were older (69.7 years versus 67.2 years, P <.0001) compared to ICH patients without MI. The in-hospital mortality was higher (40.9% versus 25.5%, p≤.0001) among ICH patients with AMI in both univariate and multivariate analysis (OR = 1.22 (1.14 -1.31), P<.0001) after adjusting for potential confounders. ICH patients with MI had higher (72.4% versus 58.8%, P <.0001) proportion of moderate to severe disability at discharge compared to ones without. Similarly, mean length of in-hospital stay (12.4 days versus 8.94 days, P <.0001) and mean hospital charges ($129328 versus $ 81984.0, P <.0001) were also higher in ICH patients with MI Conclusions: While only 3.13% of patients with ICH have an AMI, there is a 22% increase in worse outcome among those patients with AMI and ICH.

Abstract P423: Race and Ethnicity Influence Perihematomal Edema Volume in Supratentorial Intracerebral Hemorrhage

Stroke ◽  
2021 ◽  
Vol 52 (Suppl_1) ◽  
Author(s):  
Julian N Acosta ◽  
Yasheng Chen ◽  
Cameron Both ◽  
Audrey C Leasure ◽  
Fernando Testai ◽  
...  
Keyword(s):  
Intracerebral Hemorrhage ◽  
Race And Ethnicity ◽  
Learning Algorithm ◽  
Multivariable Analysis ◽  
Secondary Brain Injury ◽  
Racial Groups ◽  
Deep Learning Algorithm ◽  
Perihematomal Edema ◽  
Study Participants ◽  
Racial Ethnic

Introduction: Perihematomal Edema (PHE) is a neuroimaging biomarker of secondary brain injury in patients with spontaneous, non-traumatic intracerebral hemorrhage (ICH). There are limited data on racial/ethnic differences in the development of PHE. This dearth of data is partially driven by the time-consuming process of manually segmenting PHE. Leveraging a validated automated pipeline for PHE segmentation, we evaluated whether race and ethnicity influence baseline PHE volume in patients with ICH. Methods: The Ethnic/Racial Variations in Intracerebral Hemorrhage (ERICH) study is a prospective, multicenter study of ICH that recruited 1,000 adult participants from each of three racial/ethnic groups (non-Hispanic White, non-Hispanic Black, and Hispanic). We applied a previously validated deep learning algorithm to automatically determine PHE volumes on baseline CTs in these study participants. Quality control procedures were used to include only sufficiently accurate PHE measurements. Linear regression was used to identify factors associated with log-transformed PHE volume and to identify differences across Ethnic/Racial groups. Results: Our imaging pipeline provided good quality baseline PHE measurements on 2,008 out of 3,000 ERICH study participants. After excluding infratentorial hemorrhages (273) and those with missing or null baseline ICH volume (49), 1,686 remained for analysis (median age 59 [IQR 51-71], 687 [41%] female sex). Median PHE volume was 12.0 (IQR 4.8-27.1) for whites, 11.9 (IQR 4.5-26.1) for Hispanics and 8.3 (IQR 3.0-19.2) for blacks. Compared to Blacks, Hispanics (beta 0.22; 95%CI 0.11-0.32; p<0.001) and Whites (beta 0.20; 95%CI 0.07-0.33; p=0.003) had higher baseline PHE volumes, in multivariable analysis adjusting for age, sex, ICH location, log-baseline ICH volume, log-baseline intraventricular volume, and systolic blood pressure on admission. Conclusion: Race and ethnicity influence the volume of baseline PHE. Further studies are needed to validate our results and investigate the biological underpinnings of this difference.

Clinician Perspectives on Challenges to Patient Centered Care at the End of Life

2016 ◽  
Vol 36 (4) ◽  
pp. 401-415 ◽  
Author(s):  
Shoshana H. Bardach ◽  
Edward J. Dunn ◽  
J. Christopher Stein
Keyword(s):  
End Of Life ◽  
Medical Center ◽  
Local Policy ◽  
Do Not Resuscitate ◽  
Patient Centered ◽  
Code Status ◽  
Respiratory Therapists ◽  
Centered Care ◽  
Preventive Ethics ◽  
Dnr Orders

Discussions regarding patient preferences for resuscitation are often delayed and preferences may be neglected, leading to the receipt of unwanted medical care. To better understand barriers to the expression and realization of patients’ end of life wishes, a preventive ethics team in one Veterans Affairs Medical Center conducted a survey of physicians, nurses, social workers, and respiratory therapists. Surveys were analyzed through qualitative analysis, using sorting methodologies to identify themes. Analysis revealed barriers to patient wishes being identified and followed, including discomfort conducting end-of-life discussions, difficulty locating patients’ preferences in medical records, challenges with expiring do not resuscitate (DNR) orders, and confusion over terminology. Based on these findings, the preventive ethics team proposed new terminology for code status preferences, elimination of the local policy for expiration of DNR orders, and enhanced systems for storing and retrieving patients’ end-of-life preferences. Educational efforts were initiated to facilitate implementation of the proposed changes.

Abstract 92: Demographic, Regional and Social Predictors of Do Not Resuscitate Order Utilization in Intracerebral Hemorrhage: A National Perspective

Stroke ◽  
2015 ◽  
Vol 46 (suppl_1) ◽  
Author(s):  
Vishal B Jani ◽  
Achint Patel ◽  
Girish Nadkarni ◽  
Alexandre Benjo ◽  
Narender Annapureddy ◽  
...  
Keyword(s):  
Intracerebral Hemorrhage ◽  
Private Insurance ◽  
Female Gender ◽  
P Value ◽  
Weighted Estimate ◽  
Do Not Resuscitate ◽  
Discrimination Power ◽  
Social Predictors ◽  
North East ◽  
Dnr Orders

Background: Non-traumatic Intracerebral hemorrhage (ICH) is a life-threatening condition associated with substantial morbidity and mortality. Do-not-resuscitate (DNR) orders have recently linked to poor outcomes in ICH patients probably due to the inactive management associated with these orders. Hypothesis: We tested the hypothesis that demographic, regional and social factors not related to ICH severity are significant predictors of DNR utilization. Methods: We reviewed the Healthcare Cost and Utilization Project's Nationwide Inpatient Sample (NIS) database from 2011-2012 for ICH admissions in patients >18 years using the ICD 9-CM code 431. We defined DNR status with ICD code - V49.86 entered during the same admission as a secondary diagnosis and estimated severity of illness by the 3M™ All Patient Refined DRG (APR DRG) classification System. A hierarchical two level multivariate regression model were generated to estimate odds ratios (OR) for predictors of DNR utilization and discrimination power of models was assessed using C statistics. We considered a two tailed p value of <0.01 to be significant. Results: We analyzed 25768 pts (weighted estimate 126254) with ICH out of which 4620 (18%) pts (weighted estimate 22668) had DNR orders placed. In multivariable regression analysis, female gender (OR 1.2, 95% CI 1.2-1.3), Ethnicity [White(OR 1.6, 95% CI 1.5-1.7) and Hispanic(OR 1.2, 95% CI 1.1-1.3) compared to Black], Insurance [Medicare (OR 1.1, 95% CI 1.1-1.2) and self or no pay (OR 1.1, 95% CI 1.0-1.2) compared to private insurance], Hospital location [West (OR 1.6, 95% CI 1.2-2.1) compared to North-East ] were significantly associated with high DNR utilization rates after adjusting for patient level, hospital level characteristics, APR DRG severity scale and other clinical characteristics. Conclusions: In conclusion, demographic (female gender/ethnicity), social (insurance status) and regional (hospital location) are significantly associated with increased DNR utilization. The reasons for this are likely multifactorial, qualitative, linked to both patient and provider practices and need to be explored in more detail.

Abstract 49: Withdrawal of Care in the Ethnic/Racial Variations of Intracerebral Hemorrhage (ERICH) Study

Stroke ◽  
2013 ◽  
Vol 44 (suppl_1) ◽  
Author(s):  
Kevin N Sheth ◽  
Sharyl R Martini ◽  
David L Tirschwell ◽  
Kyra J Becker ◽  
Bradford B Worrall ◽  
...  
Keyword(s):  
Intracerebral Hemorrhage ◽  
Ethnic Groups ◽  
Minority Groups ◽  
Do Not Resuscitate ◽  
Racial Variations ◽  
Baseline Characteristics ◽  
Race Ethnicity ◽  
Gcs Score ◽  
Racial Ethnic ◽  
Hispanic White

Introduction: Withdrawal of care (WOC) during hospitalization is the most common cause of death after intracerebral hemorrhage (ICH). Prior work suggests minority groups are less likely to choose WOC. Our goal was to evaluate for differences in rates of WOC among racial/ethnic groups from the ERICH cohort. Methods: ERICH is an ongoing multicenter study of genetic and environmental risk factors for spontaneous ICH. We analyzed data from the first 725 individuals. Baseline characteristics,do not resuscitate (DNR) status, intensive care procedures, and WOC were prospectively recorded. A central core analyzed all imaging. We compared characteristics among patients with and without eventual WOC and by race/ethnicity. Logistic regression was used to identify variables independently associated with WOC and associations are presented as the odds ratio (95% confidence interval). Results: 9.9% (72/725) of patients underwent WOC. After controlling for age, ICH volume, initial Glasgow Coma Scale (GCS) score, and presence of intraventricular hemorrhage (IVH), there were no significant differences in WOC between non-Hispanic white, non-Hispanic black (OR 1.82; CI 0.78-4.25), and Hispanic (OR 2.16; CI 0.93-5.00) patients. There were also no differences in rates of DNR/DNI status across racial/ethnic groups. In multivariate analysis, patients who underwent WOC had larger ICH volume (1.75; 1.13-2.73); were older (1.43; 1.27-1.61), more likely to have IVH (3.21; 1.53-6.73), and had lower GCS (2.41; 1.63-3.56). While patients who underwent WOC were more likely to have a DNR/DNI order (12.7; 4.69-34.7), intubated patients were more likely to undergo WOC (4.09; 1.08-9.25), even after adjusting for ICH severity. Conclusions: In our cohort, we were able to model ICH severity and factors predictive of WOC. There were not significant racial/ethnic differences in WOC rates. Intubated patients are more likely to undergo care limitations, independent of ICH severity.

Abstract 19627: A National Perspective of the Influence of Do-Not-Resuscitate Orders on In-Hospital Mortality and Adverse Outcome after Intracerebral Hemorrhage

Circulation ◽  
2014 ◽  
Vol 130 (suppl_2) ◽  
Author(s):  
Achint Patel ◽  
Sopan Lahewala ◽  
Neil Patel ◽  
Girish N Nadkarni ◽  
Grishma Dhaduk ◽  
...  
Keyword(s):  
Intracerebral Hemorrhage ◽  
Hospital Mortality ◽  
Adverse Outcome ◽  
Adverse Outcomes ◽  
Categorical Variables ◽  
Do Not Resuscitate ◽  
Multilevel Regression ◽  
Co Morbidity ◽  
The Impact ◽  
Dnr Orders

Background: The impact of do-not-resuscitate (DNR) orders on outcomes has not been systematically evaluated in Intracerebral Hemorrhage (ICH). Hypothesis: We assessed the impact of DNR orders in ICH and its association to mortality/related adverse outcomes. Methods: We reviewed the Healthcare Cost and Utilization Project’s Nationwide Inpatient Sample(NIS) database of 2011 for ICH using ICD 9-CM codes(431).This represents 20% of all US hospital patients and weighted numbers represent national estimates. We defined patients’ DNR status with ICD code V49.86 and comorbid conditions by Deyo’s modification of Charlson’s Comorbidity Index (CCI). We only included adult patients in our analysis. Our primary outcomes of interest were in-hospital mortality and adverse outcome (composite of mortality & discharge other than home). We utilized chi-square test for univariable analysis for categorical variables and generated hierarchical multilevel regression models to determine independent predictors of mortality and adverse outcome. Results: We analyzed a total of 13440 pts (weighted n= 64617) with ICH of which 2029 (weighted n=9713) patients had DNR status. The proportions of mortality (56% vs. 19%, p<0.001) and adverse outcome(89% vs 70%) were higher in patients with DNR orders. Even after adjusting for confounders (demographics, Deyo’s modification of charlson’s co-morbidity index, admission type (elective vs emergent), hospital region, hospital teaching status, hospital ICH volume and primary payer), DNR status was associated with higher in hospital mortality (OR 6.98, 95% CI 6.58-7.41), p<0.001) and higher odds of adverse outcome (OR 3.98, 95% CI 3.64-4.34, p<0.001). Conclusion: DNR status in patients admitted with ICH appears to be a independent and significant predictor of substantially increased hospital mortality and adverse outcomes. The reasons for this are multifactorial and likely involve patient as well as systematic factors.Further studies including both quantitative/qualitative aspects are warranted to investigate these factors in detail.

“Allow natural death” versus “Do not resuscitate”: What do patients with advanced cancer choose?

2014 ◽  
Vol 32 (31_suppl) ◽  
pp. 89-89
Author(s):  
Milos Miljkovic ◽  
Dennis Omoding Emuron ◽  
Lori Rhodes ◽  
Joseph Abraham ◽  
Kenneth David Miller
Keyword(s):  
End Of Life ◽  
Advanced Cancer ◽  
Cancer Education ◽  
Living Will ◽  
Tube Placement ◽  
Similar Proportion ◽  
Do Not Resuscitate ◽  
Natural Death ◽  
Code Status ◽  
Dnr Orders

89 Background: Many patients with advanced cancer at our hospital request full resuscitative efforts at the end of life. In the first in a series of quality improvement projects to improve end-of-life (EOL) care, we assessed the knowledge and attitudes of patients towards it to determine if “Allow Natural Death” (AND) orders were more acceptable than “Do Not Resuscitate” (DNR) orders. Methods: Adult patients with advanced cancer being treated at a single community hospital were invited to participate. The first 100 consenting patients were surveyed regarding their diagnosis, prognosis, and attitudes about critical care and resuscitation. They were then presented with hypothetical scenarios in which a decision on their code status had to be made if they had 1 year, 6 months or 1 month left to live. Fifty patients were given a choice between being “full code“ and “DNR”, and 50 could choose between ”full code" and “AND”. Results: Participants were equally likely to choose either of the “no code” options in all hypothetical scenarios (p > 0.54). The choice was not affected by age, sex, race, type of cancer, education, or income level. Patients who said they would want life-prolonging measures such as CPR, tracheostomies, and feeding tube placement in case of a permanent vegetative state were significantly less likely to choose “AND” than “full code” (p=0.001–0.002). A similar proportion of patients who had a living will chose “AND” and “DNR” orders instead of “full code” in all the scenarios (47–74% and 63–71%). In contrast, among patients who did not have a living will 52% chose “DNR”, while 19% opted for “AND”. More than a third (39 of 93) patients were not aware their illness was terminal. Conclusions: We hypothesized that “AND” orders may be more acceptable to patients with advanced cancer, but the wording of the “no code” order does not seem to be related to the patients’ code status decisions. The “Do not resuscitate” phrasing may be more acceptable to patients who view life-prolonging measures favorably.

Differential Trends in Racial Preferences for Cosmetic Surgery Procedures

2011 ◽  
Vol 77 (8) ◽  
pp. 1081-1085 ◽  
Author(s):  
Tahira I. Prendergast ◽  
Sharon K. Ong'Uti ◽  
Gezzer Ortega ◽  
Amal L. Khoury ◽  
Ekene Onwuka ◽  
...  
Keyword(s):  
Native American ◽  
Cosmetic Surgery ◽  
Cosmetic Procedure ◽  
International Classification Of Disease ◽  
Racial Preferences ◽  
Modification Procedure ◽  
Procedure Codes ◽  
Racial Ethnic ◽  
Zip Code

There appears to be an increasing acceptance of cosmetic surgery procedures among minority populations in America. Our objective was to determine trends in elective cosmetic procedure utilization as they apply to racial/ethnic differences. A retrospective analysis was performed using the Nationwide Inpatient Sample. Adult patients undergoing elective cosmetic procedures defined by the appropriate International Classification of Disease 9 Clinical Modification procedure codes were included. Demographic characteristics and hospital course particulars were evaluated. There were 71,775 patients meeting the inclusion criteria. Median age was 48 years. The majority were female (90%), and white (65%). The median household income for the patient's zip code was most commonly in the highest economic quartile (4th quartile, 40%). The most common cosmetic procedure was liposuction (67%). The overall mean percentage change in the frequency of these procedures showed a 1.8 per cent decline among white patients, whereas Black, Hispanic, Asian, and Native American patients had an increase of 7.5 per cent, 4.7 per cent, 14.5 per cent, and 105.5 per cent, respectively. We conclude that there is an identified increasing trend in the proportion of racial/ethnic minorities represented among the recipients of cosmetic surgery procedures.

scholarly journals Treatment Restrictions and the Risk of Death in Patients With Ischemic Stroke or Intracerebral Hemorrhage

Stroke ◽  
2020 ◽  
Vol 51 (9) ◽  
pp. 2683-2689
Author(s):  
Hendrik Reinink ◽  
Burak Konya ◽  
Marjolein Geurts ◽  
L. Jaap Kappelle ◽  
H. Bart van der Worp
Keyword(s):  
Ischemic Stroke ◽  
Intracerebral Hemorrhage ◽  
Median Time ◽  
Cox Regression ◽  
University Hospital ◽  
Do Not Resuscitate ◽  
Risk Of Death ◽  
Increased Risk ◽  
Treatment Restrictions ◽  
Dnr Orders

Background and Purpose: Do-not-resuscitate (DNR) orders in the first 24 hours after intracerebral hemorrhage have been associated with an increased risk of early death. This relationship is less certain for ischemic stroke. We assessed the relation between treatment restrictions and mortality in patients with ischemic stroke and in patients with intracerebral hemorrhage. We focused on the timing of treatment restrictions after admission and the type of treatment restriction (DNR order versus more restrictive care). Methods: We retrospectively assessed demographic and clinical data, timing and type of treatment restrictions, and vital status at 3 months for 622 consecutive stroke patients primarily admitted to a Dutch university hospital. We used a Cox regression model, with adjustment for age, sex, comorbidities, and stroke type and severity. Results: Treatment restrictions were installed in 226 (36%) patients, more frequently after intracerebral hemorrhage (51%) than after ischemic stroke (32%). In 187 patients (83%), these were installed in the first 24 hours. Treatment restrictions installed within the first 24 hours after hospital admission and those installed later were independently associated with death at 90 days (adjusted hazard ratios, 5.41 [95% CI, 3.17–9.22] and 5.36 [95% CI, 2.20–13.05], respectively). Statistically significant associations were also found in patients with ischemic stroke and in patients with just an early DNR order. In those who died, the median time between a DNR order and death was 520 hours (interquartile range, 53–737). Conclusions: The strong relation between treatment restrictions (including DNR orders) and death and the long median time between a DNR order and death suggest that this relation may, in part, be causal, possibly due to an overall lack of aggressive care.

Sign in / Sign up

Export Citation Format

Share Document