Research

Public Health Reports

2017, Vol. 132(6) 669-675
Death Certification Errors and the Effect ª 2017, Association of Schools and
Programs of Public Health

on Mortality Statistics All rights reserved.

Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0033354917736514
journals.sagepub.com/home/phr

Lauri McGivern, MPH1, Leanne Shulman, MS1,

Jan K. Carney, MD, MPH2, Steven Shapiro, MD1,
and Elizabeth Bundock, MD, PhD1

Abstract
Objective: Errors in cause and manner of death on death certificates are common and affect families, mortality statistics, and
public health research. The primary objective of this study was to characterize errors in the cause and manner of death on
death certificates completed by non–Medical Examiners. A secondary objective was to determine the effects of errors on
national mortality statistics.
Methods: We retrospectively compared 601 death certificates completed between July 1, 2015, and January 31, 2016, from
the Vermont Electronic Death Registration System with clinical summaries from medical records. Medical Examiners, blinded
to original certificates, reviewed summaries, generated mock certificates, and compared mock certificates with original cer-
tificates. They then graded errors using a scale from 1 to 4 (higher numbers indicated increased impact on interpretation of the
cause) to determine the prevalence of minor and major errors. They also compared International Classification of Diseases, 10th
Revision (ICD-10) codes on original certificates with those on mock certificates.
Results: Of 601 original death certificates, 319 (53%) had errors; 305 (51%) had major errors; and 59 (10%) had minor errors.
We found no significant differences by certifier type (physician vs nonphysician). We did find significant differences in major errors
in place of death (P < .001). Certificates for deaths occurring in hospitals were more likely to have major errors than certificates
for deaths occurring at a private residence (59% vs 39%, P < .001). A total of 580 (93%) death certificates had a change in ICD-10
codes between the original and mock certificates, of which 348 (60%) had a change in the underlying cause-of-death code.
Conclusions: Error rates on death certificates in Vermont are high and extend to ICD-10 coding, thereby affecting national
mortality statistics. Surveillance and certifier education must expand beyond local and state efforts. Simplifying and standardizing
underlying literal text for cause of death may improve accuracy, decrease coding errors, and improve national mortality statistics.

Keywords
death certificate, errors, mortality statistics, ICD coding, cause of death

The National Center for Health Statistics (NCHS) reported
>2.5 million deaths in the United States in 2013, each of
which generated a death certificate.1 The death certificate
is a legal document that serves multiple purposes, such as
providing families with the cause and manner of death and
being required to settle estates, insurance, and other survivor
benefits.2 Death certificates have a direct effect on legal
proceedings, civil and criminal, and provide important evi-
dence in court. In civil cases, death certificates are used in
personal injury actions, health care liability claims, and
insurance and workers’ compensation claims.3
Death certificates also have far-reaching effects on fam-
ilies, education programs, health surveillance, public health
research and funding, policy, and health statistics. Death
certificates track the health of a population and guide the
allocation of resources for research and health programs.4
During Hurricane Sandy, the Red Cross used death certifi-
cates to track fatalities related to the storm to provide assis-
tance to surviving family members, including counseling,
emergency housing, and other disaster-related needs.5 The
accuracy of the death certificate is paramount, but several
1
Office of the Chief Medical Examiner, Vermont Department of Health,
Burlington, VT, USA
2
Robert Larner, MD College of Medicine, University of Vermont,
Burlington, VT, USA
Corresponding Author:
Lauri McGivern, MPH, Office of the Chief Medical Examiner, Vermont
Department of Health, 111 Colchester Ave, Baird 1, Burlington,
VT 05401, USA.
Email: lauri.mcgivern@vermont.gov
670
studies have demonstrated that errors in certification are
common.4,6-16
In New York City, inaccurate cause of death on certificates
led to inaccurate health disparity tracking. One study of death
certificates in 2008 revealed that the disparity in premature
heart disease between white and black people in New York
City was underestimated because hospitals that incorrectly
overreported premature heart disease served larger propor-
tions of white people than black people.17 In a study by Yin
et al, misclassification of colon and rectal cancer deaths on
death certificates affected estimated survival rates. The study
compared cancer site data from the California Cancer Registry
with cause of death on death certificates between 1993 and
1995 and found misclassification in 700 of 11 404 (6%) colon
cancer deaths and 1958 of 5011 (39%) rectal cancer deaths, of
which 1605 of 1958 (82%) were misclassified as colon cancer.
When deaths were reclassified correctly, the cause-specific
survival rate for colon and rectal cancer decreased.18
Physicians and other health care professionals are
responsible for completing the medical certification (ie,
cause and manner of death) on death certificates but often
lack proper training and experience to do so. Many med-
ical schools do not offer formal training on death certifi-
cation, and physicians learn through on-the-job
experience.2,9,10,12-14,19 Studies conducted on the imple-
mentation of educational programs showed a reduction
in certification error rates,2,4,9,19 but educational programs
are not routine and do not reach most certifiers (eg, phy-
sicians, advanced practice registered nurses, physician
assistants). Medical Examiners, however, receive formal
training on certification and complete death certificates as
part of their daily work.
The primary objective of this study was to characterize
errors in cause and manner of death reported on the death
certificate among non–Medical Examiners. A secondary
objective was to compare International Classification of Dis-
eases, 10th Revision (ICD-10)20 coding to analyze the effect
of errors in cause of death on death certificates on the 113
Selected Causes of Death and Rankable Causes of Death21 to
determine how these errors affect national statistics.
Methods
We randomly selected 750 cases from death certificates com-
pleted by non–Medical Examiners in the Vermont Electronic
Death Registration System (EDRS)22 between July 1, 2015,
and January 31, 2016, using a data management system called
Crystal Reports, version 11.0.23 Of the 750 cases, we completed
a retrospective review on 601 (80%) certificates that met inclu-
sion criteria (ie, non–Medical Examiner case, appropriate med-
ical information to determine cause of death, medical records
within 3 months of death, medical records received during the
study period). We had no inclusion or exclusion criteria for
place of death. We abstracted data on demographic character-
istics (name, age, date of birth, date and time of death, location
and place of death), cause and manner of death (including cause
Public Health Reports 132(6)
and contributory causes), and certifier type (physician,
advanced practice registered nurse, physician assistant) from
the original certificates to Microsoft Excel.
We requested medical records on all cases. Cases were
randomly assigned for review to 1 of 5 nationally certified
medicolegal death investigators who are registered nurses
and routinely review medical records and write reports for
the purpose of cause and manner of death determination by
the Office of the Chief Medical Examiner (OCME). Blinded
to the original certification, they reviewed the records,
abstracted and summarized the medical history and circum-
stances of death, and generated a report.
The investigator report (without patient-identifying data
or information on original certification) was submitted to 2
physician Medical Examiners for their determination of
cause, contributory cause(s), and manner of death. We
entered the Medical Examiners’ determinations into the
study database. Medical Examiners met to review cases and
develop a mock death certificate (including cause, contribu-
tory cause[s], and manner of death) by consensus. This study
was deemed not human research and considered exempt by
the University of Vermont Internal Review Board.
Statistical Analysis
We compared mock certificates with original certificates and
classified errors by using a standard error analysis form. We
analyzed errors and graded them on a scale from 1 to 4, with
4 being the most severe, based on a similar scale used to
assess certification errors in a previous Vermont univer-
sity–affiliated hospital study6:
Grade 1: Minor error (limited impact on interpretation)
1a. Inappropriate information included on death
certificate
1b. Abbreviations
Grade 2: Minor error (potential impact on interpretation)
2a. Errors of sequencing (regardless of whether an
underlying cause was provided)
2b. Multiple underlying causes of death in Part I
Grade 3: Major error (impact on interpretation of contribut-
ing causes)
3a. Major comorbidities/contributing cause(s) absent
or wrong
Grade 4: Major error (impact on interpretation of cause and
manner of death)
4a. Underlying cause listed as contributory cause (in
Part II)
4b. No underlying cause in Part I
4c(1). Wrong underlying cause
4c(2). Underlying cause not on last line of Part I
4d. Wrong manner of death
We analyzed the prevalence of each error by using 2 vari-
ables: (1) qualification of certifier (physician or nonphysician)
and (2) place of death (nonhospital facility [nursing home,
McGivern et al 671

Table 1. Errors and grades of errors on 601 randomly selected death certificates completed by non–Medical Examiners (physicians, advance
practice registered nurses, and physician assistants), Vermont, July 1, 2015, to January 31, 2016

Death Certificates With Error

Error Grades of Errora No. % (95% CI)

Any errors — 319 53 (49-57)
Major errorb 3 and 4 305 51 (47-55)
Minor errorc 1 and 2 59 10 (7-12)
Inappropriate information included 1a 21 3 (2-5)
Abbreviations used 1b 5 1 (0-2)
Errors of sequencing 2a 33 5 (4-7)
Multiple UCOD in Part Id 2b 22 4 (2-5)
Major comorbidities error 3a 232 39 (35-42)
Correct UCOD not in Part Id 4a 158 26 (23-30)
No UCOD in Part Id 4b 92 15 (12-18)
Wrong UCOD on death certificate 4c 107 18 (15-21)
UCOD not on the last lined 4c 174 29 (25-33)
Wrong manner of death 4d 22 4 (2-5)
Abbreviation: UCOD, underlying cause of death.
a
Modified from a similar scale used to assess certification errors in a previous Vermont university–affiliated hospital study.6
b
A major error is defined as having 1 of the following errors: major comorbidities/contributing causes absent or wrong, underlying cause listed as
contributory cause (in Part II), no underlying cause in Part I, wrong underlying cause, underlying cause not on last line of Part I, or wrong manner of death.
c
A minor error is defined as having 1 of the following errors: inappropriate information included, the use of abbreviations, errors of sequencing (mechanisms
and underlying cause not logically sequenced in Part I), or multiple underlying causes in Part I.
d
Part 1 of the death certificate describes the sequence of medical events leading directly to death, with the UCOD on the last completed line. Part 1 should
contain only 1 underlying cause, which is the etiologically specific disease or injury responsible for starting the lethal sequence.24

long-term care facility, or hospice], hospital [inpatient,
emergency department, outpatient, or intensive care unit],
or private residence). We used the Pearson w2 goodness-of-
fit test for uniform distribution across the categories of
each independent variable (qualification of certifier, place
of death) to determine if any 2 categories differed signif-
icantly in error prevalence. If we detected a significant
deviation from the uniform distribution of error rates,
we completed pairwise comparisons with z tests using a
Bonferroni correction for multiple comparisons. We per-
formed all analyses using Microsoft Excel, with a ¼ .05
considered significant.
We compared ICD-10 coding for the original certificates
with coding for the mock certificates in 580 of the 601 cases.
In 21 cases, the original certificate was amended by the
OCME before records were sent to NCHS for initial coding;
therefore, original coding was not available. According to
routine vital records procedures, original certifications were
coded by NCHS through ICD-10 cause-of-death lists for
tabulating mortality statistics20; codes returned from NCHS
were extracted from the Vermont EDRS. We sent the literal
text of the mock certificates to NCHS, which followed the
same ICD-10 coding protocol. NCHS returned codes for
each certificate, and we entered them in the study database.
We analyzed differences in order of codes, number of codes,
and specific codes generated from original and mock certif-
icates. We assessed the potential effect of certificate errors
on national statistics by examining how the difference in
underlying cause-of-death code affected the 113 Selected
Causes of Death list and the Rankable Causes of Death list,
which are used by NCHS to tabulate and disseminate mor-
tality statistics.21
Results
Error Analysis
Of the 601 original death certificates examined, 319
(53%) had errors; 305 (51%; 95% CI, 47%-55%) had
major errors; and 59 (10%; 95% CI, 7%-12%) had minor
errors. Most certificates with minor errors also had major
errors; only 14 certificates had minor errors but no major
errors (Table 1).
We found no significant difference between physician
and nonphysician certifiers in the percentage of certifi-
cates with major errors (51% vs 48%) or minor errors
(9% vs 13%). We found a significant difference in the
percentage of certificates with major errors (but not minor
errors) by place of death. Certificates for deaths in hos-
pitals were more likely to have major errors than certifi-
cates for deaths in a private residence (59% vs 39%, P <
.001). Certificates indicating deaths in a nonhospital facil-
ity were also more likely to have major errors than cer-
tificates indicating deaths in a private residence (53% vs
39%, P ¼ .006) (Table 2).
Before consensus, Medical Examiners were discordant
for underlying cause of death in 71 of 601 (12%) mock
certificates and for contributory causes of death in 149 of
601 (25%) mock certificates. Fifty-five percent of Medi-
cal Examiner discordance in underlying cause of death
672 Public Health Reports 132(6)

Table 2. Rates of minor and major errors by place of death in 601 randomly selected death certificates completed by non–Medical
Examiners (physicians, advanced practice registered nurses, and physician assistants), Vermont, July 1, 2015, to January 31, 2016

Death Certificates With

Errors

Error Typea: Place of Death Total No. of Death Certificates No. % (95% CI) P Valueb
Majorc <.001
Nursing home, long-term care facility, or hospice 188 100 53 (46-60)
Hospital (inpatient, emergency department, outpatient, 212 126 59 (53-66)
or intensive care unit)
Private residence 199 78 39 (32-46)
Minord .92
Nursing home, long-term care facility, or hospice 188 18 10 (5-14)
Hospital (inpatient, emergency department, outpatient, 212 20 9 (5-13)
or intensive care unit)
Private residence 199 20 11 (6-15)
a
Based on a similar scale used to assess certification errors in a previous Vermont university–affiliated hospital study.6
b
Determined by Pearson w2 goodness-of-fit test across all 3 places of death. P < .001 is considered significant.
c
A major error is defined as having 1 of the following errors: major comorbidities/contributing causes absent or wrong, underlying cause listed as
contributory cause (in Part II), no underlying cause in Part I, wrong underlying cause, underlying cause not on last line of Part I, or wrong manner of death.
d
A minor error is defined as having 1 of the following errors: the use of abbreviations, errors of sequencing (mechanisms and underlying cause not logically
sequenced in Part I), or multiple underlying causes in Part I.

Table 3. Change in UCOD ICD-10 code affecting the 113 Selected Causes of Deatha and the Rankable Causes of Deathb after comparison
of original and mock death certificates,c Vermont, July 1, 2015, to January 31, 2016

Death Certificates

Type of Change in UCOD ICD-10 Code Total No. No. (%) With Change
Any change 580 348 (60)
UCOD code change that affects the 113 Selected Causes of Death 348 289 (83)
UCOD code change that affects the Rankable Causes of Death 348 169 (49)
Abbreviations: ICD-10, International Classification of Diseases, 10th Revision; UCOD, underlying cause of death.
a
The 113 Selected Causes of Death list is used for the general analysis of mortality and for ranking the leading causes of death.21
b
The Rankable Causes of Death list is a method used to present mortality statistics. Cause-of-death ranking is useful in illustrating the relative burden of cause-
specific mortality.21
c
Medical Examiners, blinded to original death certificates, reviewed medical record summaries and determined the cause, contributory cause, and manner of
death, creating a mock death certificate that was later compared with the original death certificate.

was due to inversion of cause and contributory cause.
Therefore, notable discordance due to differences of med-
ical opinion occurred in only 32 of 601 (5%) mock
certificates.
Coding Comparison
Of 580 death certificates, 537 (93%) had a difference in ICD-
10 codes between the original certificates and mock certifi-
cates, and 348 (60%) had a change in the underlying cause-
of-death code (Table 3). Of these 348 cases, 231 (66%) had at
least 1 type of error, most commonly a grade 4 error (221 of
231); the other 117 cases did not have a certification error but
did have a change in underlying cause-of-death ICD-10 code.
Of the 117 certificates without a certification error, 80 (68%)
of the underlying cause-of-death ICD-10 code changes were
associated with a variation in literal text between original
certificates and mock certificates; 18 (15%) were caused
by inconsistent coding; 13 (11%) were caused by cause of
death and contributory cause of death being reversed
between original certificates and mock certificates; 3 (3%)
had a mechanism coded; and 3 (3%) were not coded from the
last line. A few certificates had nearly identical original and
mock certificates but were coded differently.
Of the 348 cases with an underlying cause-of-death code
change, 289 (83%) would have affected the 113 Selected
Causes of Death list, meaning that the change in code
resulted in the death being categorized as a different cause
according to the 113 Selected Causes of Death groupings
(Table 3). A total of 169 of 348 (49%) cases with an under-
lying cause-of-death ICD-10 code change would have
changed categories in the Rankable Causes of Death list.
We examined how various types of literal text errors in the
underlying cause of death affect the 113 Selected Causes of
Death and rankable cause categories (Table 3). Absence of
the correct underlying cause of death in Part I (141 certifi-
cates) resulted in a category change in the 113 Selected
Causes of Death list in 115 (82%) certificates and a category
McGivern et al
change in the Rankable Causes of Death list in 90 (64%)
certificates. Ninety-one certificates did not have the correct
underlying cause of death anywhere on the certificate, which
resulted in a change in the 113 Selected Causes of Death list
in 79 (87%) certificates and a category change in the Rank-
able Causes of Death list in 57 (63%) certificates. A total of
156 certificates did not have the underlying cause of death as
the last line of Part I, which resulted in a category change in
the 113 Selected Causes of Death list in 128 (82%) certifi-
cates and a category change in the Rankable Causes of Death
list in 99 (63%) certificates.
Discussion
Our study, which used Medical Examiner determinations
based on medical records as the gold standard, indicated a
large percentage of errors in death certificates completed by
non-Medical Examiners. Error rates between physician and
nonphysician certifiers were not significantly different. This
finding was not surprising, because neither group receives
formal training in certification and both groups have the
same requirement to read an abbreviated tutorial during reg-
istration for use of the Vermont EDRS.
Previous studies of death certification errors did not use
clearly defined rating scales, making it difficult to compare
them with our study. Our finding that death certificates for
hospital deaths had a higher percentage of major errors
(59%) compared with previous studies in the same setting
that reported rates of 24% to 45%.4,6,13,15 Some discrepancy
in percentage of major errors was likely caused by differing
methodologies and definitions of major errors. For example,
Myers et al4 found major errors (defined as no underlying
cause of death, sequencing errors or no underlying cause of
death on the last line, or multiple competing causes) on 48 of
146 (32%) certificates issued by internal medicine residents.
If we had used a similar definition, the percentage of errors in
our study would have been comparable. The methodology
and error classification scheme in Pritt et al was the closest to
the one we used, but they did not classify errors in the con-
tributory cause as major. They found that 34% of death cer-
tificates reviewed had a major error.6 Our study classified
absent or wrong major contributory causes of death as a
major error, which we felt was important because contribu-
tory causes of death are used in algorithms to determine
underlying causes of death. The leading preventable causes
of death, such as obesity and diabetes, are often contributory
causes of death,24-26 and using this information provides a
more accurate, literature-based, and public health perspec-
tive. If major error was redefined as only grade 4 errors
(wrong or absent underlying cause of death in Part I, under-
lying cause of death not on last line, or wrong manner), then
the percentage of major errors in our study would be 30%.
Overall, we found significantly fewer minor errors than
major errors (10% vs 51%). The Vermont EDRS prompts
certifiers to avoid the use of abbreviations, but other minor
errors tabulated in this study (eg, inappropriate information,
673
errors of sequencing, and multiple causes in Part I) are not
detected by Vermont EDRS prompts. We classified errors of
sequencing and multiple underlying causes in Part I as minor
errors because they were expected to have only a minimal
effect on the overall interpretation of the cause from reading
the literal text. But, from a coding standpoint, the errors
potentially have a significant effect. The most egregious
sequencing error, not having the underlying cause of death
as the last line, was therefore considered a major error.
Indeed, when this error type was present, it most often caused
a category change in the 113 Selected Causes of Death and
Rankable Causes of Death lists.
Inaccuracies in death certification are common, but our
findings indicate that Vermont has a similar or higher per-
centage of major errors compared with other regions as
reported in the literature.4,6,13,15 The OCME takes an active
role in reviewing all death certificates and providing feed-
back to certifiers either through email or a telephone call.
The OCME offers education to medical students, pathology
residents, and interns on proper death certification. Ver-
mont’s EDRS requires first-time users to complete a brief
tutorial about death certification. EDRS provides feedback to
certifiers through “soft” and “hard” edits when information is
missing or a cause is nonspecific or suggests a nonnatural
death. The edits require the certifier to stop and review the
cause of death for accuracy, completeness, and Medical
Examiner consideration. Despite these efforts, the error rate
is still high.
The potential impact of certification errors and changes in
ICD-10 underlying cause-of-death codes on mortality statis-
tics was large and unexpected. The underlying cause of death
changed in 60% of certificates. Eighty-three percent of these
changes would have affected the 113 Selected Causes of
Death list, and almost half would have affected the Rankable
Causes of Death list. Although certification errors and varia-
tions in certification style were the primary reasons for code
changes, the coding process clearly contributed. In 34% of
cases in which the ICD-10 code for underlying cause of death
changed, our grading scheme detected no error of certifica-
tion. The reason for code change was most frequently varia-
tion in literal text that affected nosologists’ interpretation.
Accurate certification is the foundation for good epidemio-
logic data, but the certification error rate is not directly pro-
portional to the error of mortality statistics, because
nosologist interpretation and coding algorithms (manual vs
automated coding) may mitigate some certification errors
and introduce other errors. A more in-depth analysis of the
effect of certification errors in the context of coding algo-
rithms is needed.
Strengths
This study was unique for several reasons. The use of orig-
inal medical records rather than coded data or problem lists
and the use of physician reviewers with expertise in death
certification provided optimal retrospective evaluation. We
674
used a clearly defined error scale, which can be adjusted to
allow for comparison or use in future studies. In Vermont, all
death certificates are entered into the EDRS and reviewed by
the OCME as part of routine death surveillance and improve-
ment of certificate quality and completeness. Many Medical
Examiner and coroner offices lack the personnel and
resources to perform such comprehensive surveillance.
Unlike previous studies, this study was not limited to specific
populations by selection for place of death, certifier type, or
major cause category. In addition, similar studies to assess
error rates in death certification do not include the effects on
ICD-10 coding and mortality statistics.
Limitations
This study also had several limitations. Vermont is unique in
that it has 1 centralized office with the ability to review all
death certificates in the state through an electronic death
registry and either amend certificates directly or provide
feedback to certifiers to improve quality. Therefore, the
results from this study may not be generalizable to jurisdic-
tions that do not have a similar system, available personnel,
funding, or access to an electronic death registry. Certificates
for deaths in a private residence had a significantly lower
percentage of major errors (39%) than certificates for deaths
in hospital and nonhospital facilities (59% and 53%, respec-
tively). However, these findings may have been affected by
sampling bias caused by exclusion criteria and may be gen-
eralized only to cases that have adequate medical records.
Therefore, we cannot draw a strong conclusion. No studies
on death certification errors in private residences or nonhos-
pital facilities have been published, so comparisons cannot
be made.
Our determination of cause and manner of death was
limited by the quality of medical records. We excluded many
death certificates because of inadequate documentation,
which was most apparent in deaths at long-term care facili-
ties and nursing homes, where many records were handwrit-
ten and difficult to read. The high exclusion rate from
nonhospital facilities may have falsely reduced the reported
number of errors. Some discrepancies between original and
mock certifications may have been caused by differences in
medical opinion rather than overt error, particularly in multi-
factorial deaths with several major medical conditions. How-
ever, mock certification by consensus limited this sort of
error.
Conclusions
Implementation of the Vermont EDRS in 2008 has improved
timeliness and reduced errors, but Vermont still has a high
error rate despite active interventions to improve quality.
Jurisdictions with less active efforts may have even higher
error rates. Errors in certification are further compounded by
errors in coding. Death certificate surveillance must expand
beyond efforts by local vital records departments and
Public Health Reports 132(6)
Medical Examiner and coroner offices. Methods to further
improve quality may include comprehensive reviews of hos-
pitals’ own death certificates with internal feedback to certi-
fiers, similar to other ongoing quality improvement
initiatives. Standardizing or simplifying underlying cause-
of-death literal text on the death certificate may also improve
accuracy and decrease coding errors, thereby improving
national mortality statistics.
Further research is needed in several related areas.
Research on death certification errors that occur outside of
an academic setting or hospital is limited. Similar studies in
other regions of the country, with different systems, may
shed light on the extent of the problem in the United States.
In our experience, education and outreach have a negligible
effect on the quality and accuracy of cause-of-death state-
ments. The effectiveness of educational programs should
therefore be evaluated before implementation. More studies
are needed to evaluate the relative contributions of death
certification and coding errors on national mortality
statistics.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for
the research, authorship, and/or publication of this article: This
study was supported by the Vital Statistics Cooperative Program
contract between the Centers for Disease Control and Prevention
and the Vermont Department of Health. The contents of this article
are solely the responsibility of the authors and do not necessarily
represent the official views of Centers for Disease Control and
Prevention.
References
1. Xiu J, Murphy SL, Kochanek KD, Bastian BA. Deaths: final
data for 2013. Natl Vital Stat Rep. 2016;64(2):1-119.
2. Brooks EG, Reed KD. Principles and pitfalls: a guide to death
certification. Clin Med Res. 2015;13(2):74-82.
3. Thornton RG. Death certificates. Proc (Bayl Univ Med Cent).
2006;19(3):285-286.
4. Myers KA, Farquhar DR. Improving the accuracy of death
certification. CMAJ. 1998;158(10):1317-1323.
5. Casey-Lockyer M, Heick RJ, Mertzlufft CE, et al. Deaths asso-
ciated with Hurricane Sandy—October-November 2012.
MMWR Morb Mortal Wkly Rep. 2013;62(20):393-397.
6. Pritt BS, Hardin NJ, Richmond JA, Shapiro SL. Death certifi-
cation errors at an academic institution. Arch Pathol Lab Med.
2005;129(11):1476-1479.
7. Fischtein D, Cina SJ. Errors on death certificates requiring
amendments: the Broward County experience. Am J Forensic
Med Pathol. 2011;32(2):146-148.
8. Croft PR, Lathrop SL, Zumwalt RE. Amended cause and man-
ner of death certification: a six-year review of the New Mexico
experience. J Forensic Sci. 2006;51(3):651-656.
McGivern et al
9. Middleton D, Anderson R, Billingsly T, Virgil NBM, Wim-
berly Y, Lee R. Death certification: issues and interventions.
Open J Prev Med. 2011;1(3):167-170.
10. Tatsumi K, Shapiro S, Bundock E. Death certificate surveil-
lance: a component of death investigation. Acad Forensic
Pathol. 2013;3:99-104.
11. Hanzlick R. Quality assurance review of death certificates: a
pilot study. Am J Forensic Med Pathol. 2005;26(1):63-65.
12. Cambridge B, Cina SJ. The accuracy of death certificate com-
pletion in a suburban community. Am J Forensic Med Pathol.
2010;31(3):232-235.
13. Smith Sehdev AE, Hutchins GM. Problems with proper com-
pletion and accuracy of the cause-of-death statement. Arch
Intern Med. 2001;161(2):277-284.
14. Wexelman BA, Eden E, Rose KM. Survey of New York City
resident physicians on cause-of-death reporting, 2010. Prev
Chronic Dis. 2013;10:E76.
15. Cina SJ, Selby DM, Clark B. Accuracy of death certification in
two tertiary care military hospitals. Mil Med. 1999;164(12):
897-899.
16. Nashelsky MB, Lawrence CH. Accuracy of cause of death
determination without forensic autopsy examination. Am J
Forensic Medic Pathol. 2003;24(4):313-319.
17. Johns LE, Madsen AM, Maduro G, Zimmerman R, Konty K,
Begier E. A case study of the impact of inaccurate cause-of-
death reporting on health disparity tracking: New York City
premature cardiovascular mortality. Am J Public Health. 2013;
103(4):733-739.
18. Yin D, Morris CR, Bates JH, German RR. Effect of misclassi-
fied underlying cause of death on survival estimates of colon
and rectal cancer. J Natl Cancer Inst. 2011;103(14):1130-1133.
675
19. Degani AT, Patel RM, Smith BE, Grimsley E. The effect of
student training on accuracy of completion of death certificates.
Med Educ Online. 2009;14:17.
20. Centers for Disease Control and Prevention, National Center
for Health Statistics. International Classification of Diseases,
Tenth Revision (ICD-10). https://www.cdc.gov/nchs/icd/icd10.
htm. Accessed August 30, 2017.
21. Centers for Disease Control and Prevention, National Center
for Health Statistics. Instruction manual, part 9: ICD-10 cause-
of-death lists for tabulating mortality statistics (updated 2002 to
include ICD codes for terrorism deaths for data year 2001 and
WHO updates to ICD-10 for data year 2003). https://www.cdc.
gov/nchs/data/dvs/im9_2002.pdf.pdf. Published 2002.
Accessed September 6, 2016.
22. Vermont Department of Health, Office of Vital Statistics.
Vermont Electronic Death Registry. http://www.healthver
mont.gov/health-statistics-vital-records/vital-records-popula
tion-data/vital-records-reporting-edrs. Accessed August
30, 2017.
23. Business Objects Software Ltd. SAP Crystal Reports Version
11.0. Dublin, Ireland: Business Objects Software Ltd; 2004.
24. Centers for Disease Control and Prevention. CDC national
health report highlights. https://www.cdc.gov/healthreport/pub
lications/compendium.pdf. Accessed June 16, 2017.
25. Borrell LN, Samuel L. Body mass index categories and mor-
tality risk in US adults: the effect of overweight and obesity on
advancing death. Am J Public Health. 2014;104(3):512-519.
26. Danaei G, Ding EL, Mozaffarian D, et al. The preventable
causes of death in the United States: comparative risk assess-
ment of dietary, lifestyle, and metabolic risk factors. PLoS
Med. 2009;6(4):e1000058.