See it and Catch it! Recognizing the Thought Traps that Negatively Impact How...
PPT with Critical appraisal
1. The impact of postpartum hemorrhage on
hospital length of stay and inpatient mortality: a
National Inpatient Sample-based analysis
Ariela L. Marshall, MD; Urshila Durani, MD; Adam Bartley, MS; Clinton E. Hagen, MS; Aneel
Ashrani, MD, MS; Carl Rose, MD; Ronald S. Go, MD; Rajiv K. Pruthi, MBBS
3. INTRODUCTION
Postpartum
hemorrhage (PPH) is
a leading cause of
maternal morbidity
and mortality
Increased
during the
1990s and
2000s
In the US, the incidence is estimated
at approximately 3% and also
increased during 1990s to 2000s
4. The most common cause of PPH :
• Uterine atony (at least 80% of cases)
• Genital trauma (lacerations, incisions, uterine ruptu-
re, and others)
• Retained placenta
• Coagulopathy comprising other major
etiologies
INTRODUCTION
Predisposing risk factors for PPH :
• Maternal epidemiologic characteristics
(increased maternal age, parity)
• Obstetrical characteristics (gestational age,
induction and augmentation of labor,
cesarean delivery)
• Sociodemographic characteristics (race,
income, hospital type and location).
• The association of PPH with increased maternal mortality is well established.
• It has furthermore been suggested that PPH contributes disproportionately to severe maternal morbidity
• Maternal morbidity and mortality individual patients, clinicians, and health policy research and development
• Population-based observational studies to strengthen the available epidemiologic evidence regarding PPH
6. This study was exempt from institutional review board approval
given that it involved collection of existing data via a large
database in such a manner that the subjects could not be
identified at any time, directly or through identifiers linked to
the subjects.
MATERIALS AND METHODS
Institutional review board exemption
7. The NIS is a component of the Healthcare Cost and Utilization
Project (HCUP) sponsored by the Agency for Healthcare
Research and Quality.
MATERIALS AND METHODS
National Inpatient Sample
Agency for Healthcare Research and Quality. Overview of the Nationwide Inpatient Sample (NIS). Rockville (MD). Available at: https://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed March 31, 2016.
8. Beginning in 2012, the NIS underwent a redesign initiative.
The redesign was intended to allow more representative
sampling of hospital factors including urban-rural location,
teaching status, and bed size category, as well as several
patient-specific factors.
MATERIALS AND METHODS
National Inpatient Sample
Agency for Healthcare Research and Quality. Nationwide Inpatient Sample (NIS) redesign final report. Rockville (MD). Available at: https://www.hcup-us.ahrq.gov/reports/ methods/2014-04.pdf. Accessed March
31, 2016.
9. We used NIS data from 2012 through 2013.
To identify hospitalizations for childbirth, we used criteria
demonstrated in a previous study.
To identify PPH we searched for International Classification of
Diseases, Ninth Revision (ICD-9), Clinical Modification codes
666.0-666.
MATERIALS AND METHODS
Patient identification
Kuklina EV, Whiteman MK, Hillis SD, et al. An enhanced method for identifying obstetric deliveries: implications for estimating maternal morbidity. Matern Child Health J 2008;12: 469-77.
10. Patients with 1 PPH code Were further categorized as atonic or
nonatonic.
We also identified patients who received different transfusions
using ICD-9 codes 99.03 (whole blood), 99.04 (packed cells),
99.05 (platelets), and 99.06 (coagulation factors).
Multiparous deliveries were identified by ICD-9 codes 651.0-
651.9 (excluding 651.7).
MATERIALS AND METHODS
Patient identification
Kramer MS, Berg C, Abenhaim H, et al. Incidence, risk factors, and temporal trends in severe postpartum hemorrhage. Am J Obstet Gynecol 2013;209:449.e1-7.
11. All analyses were performed using software (SAS, Version
9.4, SAS Institute, Cary, NC).
MATERIALS AND METHODS
Statistical methods
t test or X²
tests
pairwise
comparisons
to compare characteristics between PPH
and non-PPH women in the sample
significant
Sharpe D. Your chi-square test is statistically significant: now what? Practical assessment, research and evaluation. Available at: http://pareonline.net/getvn.asp?v¼20&n¼8. Accessed Oct. 10, 2016.
12. The clinical outcomes analyses took the sampling design into
consideration using SAS PROC SURVEYREG and PROC SURVEYFREQ
procedures.
MATERIALS AND METHODS
Statistical methods
We analyzed age, race, income (by ZIP code quartiles), insurance
status, and hospital type/location.
Comparisons were made by calculating the risk difference (95%
confidence interval [CI]).
LOS was summarized by mean (95% CI), with comparisons being made
by mean difference (95% CI).
13. In addition to comparing LOS and mortality, we assessed whether there
were any trends over time using earlier years of the NIS.
MATERIALS AND METHODS
Statistical methods
Trend weights provided by HCUP were used with data from 2004
through 2011.
We plotted summary statistics for both clinical outcomes over a 10-year
period.
15. NIS data from 2012 through 2013 contained
1,470,371 hospitalizations for
birth that met inclusion criteria. Using
complete-case analysis, we dropped 8%
of our sample; the final analysis data set
included 1,352,691 hospitalizations
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
16. The most common classification of PPH was uterine atony (76.6%). Other
classifications included retained placenta (8.4%), delayed/secondary (8.1%),
and coagulation defects (7.4%)
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
18. • Multiparity was more likely in both atonic and nonatonic PPH vs non-PPH
deliveries
• Hysterectomy was significantly more frequent in nonatonic PPH compared
to atonic PPH and non-PPH deliveries.
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
19. From Tabel 2 :
• Further analysis regarding race demonstrated that atonic PPH was more likely in
Hispanic women, while nonatonic PPH and non-PPH deliveries were more common
in White women.
• While the X2 test was significant for comparison of insurance, subsequent analyses
did not suggest there were any meaningful differences in any pairwise
comparisons.
• Conversely, there were significant differences for all pairwise comparisons of
hospital type, hysterectomy, and all transfusion types, excepting whole blood
products.
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
20. • Hospital LOS was greatest for non atonic PPH hospitalizations (3.67 days), followed
by atonic PH (2.98 days). LOS was shortest in deliveries not complicated by PPH 2.63
days).
• Nonatonic PPH LOS was 1.04 days longer on average than deliveries not complicated
by PPH (95% CI, 0.93-1.14; P <.001).
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
21. • After adjusting for race, multiparity, and hospital type, the
average difference was 0.97 days (P <.001).
• Atonic PPH LOS was 0.35 days longer on average than deliveries
not complicated by PPH (95% CI, 0.31-0.39; P <.001)
• After adjusting forr ace, multiparity, and hospital type, the
average difference was 0.26 days (P <.001). Using the trend
weights for earlier years, hospital LOS from 2004 through 2013
is demonstrated in Figure 1.
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
24. • In-hospital mortality was rare. There were 58 deaths recorded; 42 among
deliveries not complicated by PPH (3.2 per 100,000), 6 among deliveries
complicated by atonic PPH (19.0 per 100,000), and 10 among deliveries
complicated by non atonic PPH (103.5 per 100,000).
• Of in-hospital mortality events during the study period, 27.6%
• The risk difference between atonic PPH and non-PPH deliveries was 0.016%
(95% CI, 0.001-0.031; P ¼.04).
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
25. • There was almost no change in the difference after adjusting for race,
multiparity, and hospital type (Table 3).
• The risk difference between nonatonic PPH and non-PPH deliveries was
0.1% (95% CI, 0.036-0.164; P ¼ .002).
• Similarly, adjusting for race, multiparity, and hospital type did not affect the
risk difference (0.099%; P ¼.002) (Table 3).
• Using the trend weights for earlier years, inhospital mortality from 2004
through 2013 is shown in Figure 2.
SOCIODEMOGRAPHIC
CHARACTERISTICS
HOSPITAL LOS IN-HOSPITAL MORTALITY
28. The association between PPH and hospital LOS and inpatient
mortality ??
COMMENT
PPH was associated with a
small but statistically longer
hospital LOS
PPH was also associated with
increased mortality
*This is one of the first studies to focus on hospital LOS in addition to maternal mortality
29. Association between PPH and maternal mortality was not as
robust as suggested by prior studies
COMMENT
• The low mortality rate across the entire sample
• Possibly reflecting recent improvements in care of women with
PPH in the United States
• The prior studies utilized data derived from earlier time periods
• The majority of PPH cases were atonic but the majority of
maternal deaths due to PPH were nonatonic
30. Very Large sample size provided by NIS
this database is representative of current peripartum data, including PPH across the USA
likely generalizable to the US population as a whole
The recent timing of the available data (2012/13)
allows for accurate analysis of recent temporal developments in PPH morbidity and
mortality.
This is one of the first studies to examine the impact of PPH on inpatient
LOS in addition to mortality
therefore provides additional novel data for clini-cians, health economists, and
specialists in health care policy
Important Strength of this Study
COMMENT
31. Patients are included on the basis of a discharge diagnosis,
subject to error and / or misclassification bias. Such biases may vary by geographic
location and type of hospital
There is no uniform definition of PPH, uterine atony, coagulopathy, or
any other classification category necessary for this study
The classification and recording of transfusions of blood products also
can not be done uniformly
Limitation of The Study
COMMENT
32. Many professional societies have published guidelines for the
prevention and management of PPH
COMMENT
Using of such guidelines and protocols successfully reduced incidence of
PPH
33. COMMENT
PPH contribute to maternal death and
disability in all regions of the world
PPH was associated with several longer
hospital LOS
Maternal and neonatal hospital stays
account for >20% of all inpatient
hospitalizations in the United States
35. Further studies could focus on specific determinants of PPH incidence and
complications and tailor interventions to geographic areas, patient
populations, and hospital types with the greatest need for intervention to
reduce the adverse consequences of this potentially preventable
pregnancy complication.
Further studies...
COMMENT
37. Women with PPH experienced significantly longer LOS and
higher inpatient mortality rates than women without PPH
CONCLUSION
38. CONCLUSION
• The preponderance of these differences were attributable to nonatonic causes of PPH.
• While the association between PPH and increased mortality risk was previously reported, this is
one of the first studies to also report the association between PPH and increased hospital LOS.
• LOS, in addition to mortality, is an important area of focus for clinicians and health care policy
specialists.
• Interventions to reduce mortality and morbidity (including increased LOS) related to PPH may
allow for delivery of more cost-effective care as well as overall improvements in maternal and
population health.
40. Title of Article:
The impact of postpartum hemorrhage on hospital length of stay and inpatient
mortality: a National Inpatient Sample-based analysis
Journal :
American Journal of Obstetric and Gynecology
PICO-VIA
45. VALIDITY APPLICABLEIMPORTANCE
This study is Valid
The population in this study can represent general population because
the sample used National data from NIS. This study also used appropriate
design and analysis method to compare the variable.
46. VALIDITY APPLICABLEIMPORTANCE
This results of this study is important because it provides information for
clinician to about the need of interventions to reduce morbidity and
mortality related to postpartum hemorrhage.
47. VALIDITY APPLICABLEIMPORTANCE
This results is applicable in the country where the study conducted which
is US, but, in another country with different characteristics, may have a
different results. Besides, the study conclude that “there is a need of
interventions to reduce morbidity and mortality related to postpartum
hemorrhage that may simultaneously facilitate delivery of more cost-
effective care and improve both maternal and population health” is
applicable in all the countries.
48. - Level of evidence : Level 2d i.e Historic/Retrospective control group
study
- Recommendation : B
CONCLUSION PICO-VIA