Perinatal Periods of Risk Practice Collaborative · Phase 1 results suggest some general areas for action, but are not sufficient information alone. \爀䘀漀爀䴀愀琀攀爀渀愀氀

$Page 1: Perinatal Periods of Risk Practice Collaborative · Phase 1 results suggest some general areas for action, but are not sufficient information alone. \爀䘀漀爀䴀愀琀攀爀渀愀氀$
The Perinatal Periods of Risk Approach

Phase 2 Analytic Methods

Maternal Health/

Prematurity

Maternal Care

Newborn Care

Infant Health

Preconception Health Health Behaviors Perinatal Care etc.

Prenatal Care High Risk Referral Obstetric Care etc.

Perinatal Management Neonatal Care Pediatric Surgery etc.

Sleep Position Smoking Injury Prevention etc.

Phase 1 Narrows the Choices of Action

Presenter

Presentation Notes

Phase 1 results suggest some general areas for action, but are not sufficient information alone. For Maternal Health and Prematurity, prevention may need to focus on preconceptional health, unintended pregnancy, smoking, drug abuse, and specialized perinatal care. For Maternal Care, prevention may need to focus on early continuous prenatal care, referral of high risk pregnancies and good medical management of diabetes, seizures, post maturity or other medical problems. For newborn care, the focus may need to be advanced neonatal care and treatment of congenital anomalies. And for infant health, communities may need to focus on SIDS prevention like sleep position or breast feeding, access to a medical home and injury prevention.

Phase 1 is NOT enough. Phase 2 analyses are REQUIRED to determine which RISK FACTORS are most important in YOUR COMMUNITY…

0.8 1.1 0.6

2.0

EXCESS Fetal-Infant

Mortality Rate =4.4

Phase 2 Analysis Plan A separate Phase 2 analysis is completed for each

population and period of risk with a large “gap”. (Urban County vs USA 2000-2002 Reference Group )

MH/P 44%

MC 13%

NC 18%

IH 25%

Excess Mortality

Phase 1 and Exercise 4

How can we determine the most effective ways to prevent excess deaths?

Phase 2 Analysis Plan Depends On: ● Phase 1 results

● Availability of data

● Community priorities and resources

● Each step depends on the result of the previous step

Protocols were developed for MH/P and IH periods. Possible strategies were outlined for MC.

Phase 2 Overview

Presenter

Presentation Notes

The Phase 2 analysis plan is not as strictly laid out as is the Phase 1 plan. There are protocols, but how much of the protocol can be followed depends on what data sources are available, on how much time and effort the community wants to invest, and on what expertise is available.

Steps for Phase 2 Analysis

1. Identify causal pathways or biologic mechanisms for excess mortality

2. Estimate prevalence of risk and preventive factors by type of mechanism

3. Estimate the impact of the risk and preventive factors.

Phase 2 Overview 7

Infant health period gap

Injuries

Car seats

Child abuse/neglect

Congenital Anomalies

SIDS

Co-sleeping

Bedding

Sleep position

Phase 1

Phase 2 Step 1

Phase 2 Step 2

Phase 2 Analysis Strategy

● Eliminate consideration of risk and preventive factors that are UNLIKELY to be contributing

● Find and target KNOWN factors that are likely to be contributing

Presenter

Presentation Notes

In most cases, factors with no disparity in prevalence can be eliminated as major contributors to the disparity in mortality. Remaining factors must be identified and then prioritized according to the potential impact of prevention strategies.


Phase 2 Analytic Methods Infant Health Period of Risk

Presenter

Presentation Notes

We will walk through these steps as they play out within each period of risk.


Step 1 “Causal pathway”

What causes of death are contributing the most to excess mortality in this risk period for this population group?

Phase 2 Overview 12

Presenter

Presentation Notes

In the green box, which cause of death contributes the most to excess mortality? SUID? Infections? Accidents? In the blue box, are very low birthweight deaths due more to excess prematurity rates, or poorer survival rates?

Phase 2 Analysis---Infant health period Count of deaths by underlying cause of death code

P219 1 Birth asphyxia, unspecified Q208 1 Other congenital malformations of cardiac chambers and connections Q232 1 Congenital mitral stenosis Q249 2 Congenital malformation of the heart, unspecified Q909 1 Down's syndrome, unspecified Q913 1 Edwards' syndrome, unspecified I400 1 Infective myocarditis J129 1 Viral pneumonia, unspecified J154 2 Pneumonia due to other streptococci J180 2 Bronchopneumonia, unspecified J189 1 Pneumonia, unspecified V486 1 Passenger injured in traffic accident V892 1 Person injured in unspecified motor-vehicle accident, traffic W08 1 Fall involving other furniture W65 1 Drowning and submersion while in bathtub W75 2 Accidental suffocation and strangulation in bed W84 1 Unspecified threat to breathing X00 1 Exposure to uncontrolled fire in building or structure X44 1 Accidental poisoning & exposure to other drugs & biological substance X91 3 Assault (homicide) by hanging, strangulation, and suffocation Y079 2 Other maltreatment syndromes By unspecified person Y20 1 Hanging, strangulation, and suffocation, undetermined intent R95 15 SIDS R99 6 Other ill-defined and unspecified causes of mortality

Presenter

Presentation Notes

Coding varies according to who does it. Physicians, are supposed to fill out this part of the death certificate, but they don’t always do it. In some locations, coroners who do this are not physicians. We recommend if possible that you ask your child death review team, or Fetal infant mortality review team or even your local coroner, for information on how deaths are coded in your location.

Phase 2 Analysis--Infant Health Period The many causes of death must be grouped.

Several proposed grouping systems, such as:

● Birth defects ● Infections ● Injuries (intentional and unintentional) ● Perinatal conditions ● SIDS/SUID (possibly suffocation)

Presenter

Presentation Notes

It often works well to categorize the underlying cause of death according to the classification system used by CDC’s Postneonatal Mortality Surveillance System which includes birth defects, infections, injuries, perinatal conditions, SIDS and other causes.(MMWR Vol. 47, No. SS-2, Pages 29-30) Coding for ICD-10 has been suggested in the guidelines.

Modified Dolfus from State Infant Mortality Committee

Modified Dolfus 7/2005 ICD-9 ICD-10

1. Prematurity and related conditions

Immaturity and preterm 765.0, 765.1 P07.0, P07.1, P07.2, P07.3 Intracranial hemorrhage (mostly IVH) 431 I61

Respiratory problems 767.0, 772.1

P10.0, P10.1, P10.2, P10.4, P10.9, P29.0, P29.1, P52.0-

P52.3, P52.9 RDS (excludes transient tachypnea) 769 P22.0, P22.8, P22.9 Interstitial emphysema 770.2 P25 Pulmonary hemorrhage 770.3 P26

Chronic respiratory disease 770.7 P27 Atelectisis 770.4, 770.5 P28.0, P28.1

Other respiratory conditions 770.8 P28.2-P28.9 NEC 777.5 P77 Other specified perinatal conditions 779.8 P96.8

Presenter

Presentation Notes




2. Congenital anomaly 740-759 Q00-Q99 3. SIDS and SUID

SIDS 798 R95 Unknown 799 R99

(Decided to combine this category because of changes in reporting in recent years)

4. Obstetric Conditions PROM and incompetent cervix 761.0, 761.1 P01.0, P01.1 Multiple pregnancies 761.5 P01.5 Placental abnormalities 762.0- 762.2 P02.0, P02.1, P02.2

Presenter

Presentation Notes




5. Birth Asphyxia

Unspecified fetal distress 768.4 P20.9

Severe & unspecified birth asphyxia 768.5, 768.9 P21.0, P21.9

6. Perinatal Infections Streptococcous meningitis 320.2 G00.2

Maternal infection 760.1, 760.2,

760.8 P00.1, P00.2, P00.8

Chorioamnionitis 762.7 P02.7 Other perinatal infections 771.8 P36, P39.9

Presenter

Presentation Notes



Modified Dolfus 7/2005 ICD-9 ICD-10 7. Other infections

Infectious and parasitic diseases (we also included neonatal tetanus A33 or 771.3 in this category unlike Dolfus) 001-139 A00-B99

Meningitis

320.0, 320.1, 320.8, 320.9,

322.9 G00.0, G00.1, G00.8, G00.9,

G03.9

Respiratory infections 460-466, 480-

487, 490 J00-J06, J10-J21, J40 8. External Causes/Injuries E800-E999 V01-Y89

WE decided to include all external causes. Dolfus excluded MVAs due to railroad accidents and other than typical modes of transportation, as well as by firearms.

Presenter

Presentation Notes


Example : Child Death Review Teams frequently reclassified ● Death Cert R99 Ill-defined →as R95 SIDS

● Death Cert R95 SIDS →as W75 suffocation FIMR, medical examiners, and others may also be good sources

Death Certificate recording is not always correct. Check local practices.

1. Decide how to categorize causes of death

2. Categorize the same way for study and

reference groups

3. Count number of deaths in each category

4. Calculate “cause-specific mortality rates”

5. Compare study and reference groups rates

Cause of Death Analysis

Presenter

Presentation Notes

This is how you would do the first step of Phase 2 analysis for the Infant Health period.

Example, Cause of Death Identify causal pathways or biologic mechanisms for excess mortality

Postneonatal Deaths >1500 g State Ref Study Pop Perinatal conditions 8 1 Congenital anomalies 55 6 Infections 5 7 Injury 16 15 SIDS 61 15 Ill-defined 9 6 Other 29 4 Total Deaths 183 54

Ideal denominator is the population at risk Live Births >1500 grams, surviving 28 days

Presenter

Presentation Notes

Example thanks to Shira Rutman, Urban Indian Health Institute

Phase 2 Analysis-Infant health period Identify causal pathways or biologic mechanisms for excess mortality

In each Study and Reference Population:

CSMR=Cause-specific mortality rate

the number of IH deaths due to a specific cause number of all IH live births

Excess CSMR=Excess cause-specific mortality rate

Study Pop. CSMR – Reference Pop. CSMR

Denominator is Infant Health live births weighing >1500 grams at birth, and surviving 28 or more days

Presenter

Presentation Notes

The contribution of each cause category to the excess Infant Health mortality for each population can be calculated. The cause-specific excess mortality rate is the cause-specific rate for the study population minus the cause-specific mortality rate for the reference population. The contribution is calculated by dividing the cause-specific excess mortality rate by the total excess mortality rate for the period. The percentage of deaths by cause of death should not be used for comparison because these percentages do not take into account differences in overall mortality rates. For example, the percentages by cause may be the same for two populations while the mortality rate for is one is twice to three times the mortality rate for the other. (Fetal deaths are not included in the infant health period analysis)

Phase 2 Analysis - Infant Health Period Identify causal pathways or biologic mechanisms for excess mortality

Postneonatal Deaths >1500 g

State Ref Rate

Study Pop Rate

Excess Rate

Perinatal conditions 0.05 0.12 0.07 Congenital anomalies 0.35 0.70 0.35 Infections 0.03 0.82 0.79 Injury 0.10 1.75 1.65 SIDS 0.38 1.75 1.37 Ill-defined 0.06 0.70 0.64 Other 0.18 0.47 0.28 Total Deaths 1.15 6.32 5.16

Denominator 158,577 8,550



From CDC Wonder accessed 4/21/2011 Post-neonatal deaths among infants born at 1,500 grams or larger Black (non-Hispanic) in Philadelphia County and a Reference Group of White (non-Hispanic) mothers age 20-39, in Bucks, Chester, Delaware, and Montgomery Counties in Pennsylvania

Cause of Death Categories Black Philadelphia County

Ref Group: White age20-39 4 adjoining counties

Congenital Anomaly 10 15

Infection 4 5

Injury 12 3

Other 11 21

Perinatal Conditions 9 4

R95-SIDS 22 11

R99-Ill-defined 29 2

W75-Accidental suffocation/Strangulation in bed 0 3

Grand Total 97 64

Denominator (births>1,500 grams minus neonatal deaths) 30,187 61,823

Why so many Ill-defined for Blacks?

From CDC Wonder accessed 4/21/2011 Post-neonatal deaths among infants born at 1,500 grams or larger Black (non-Hispanic) in Philadelphia County and a Reference Group of White (non-Hispanic) mothers age 20-39, in Bucks, Chester, Delaware, and Montgomery Counties in Pennsylvania




Infection 4 5

Injury 12 3

Other 11 21


R95-SIDS 22 11



Grand Total 97 64


REALLY no suffocations?

From CDC Wonder accessed 4/21/2011 Post-neonatal deaths among infants born at 1,500 grams or larger Black (non-Hispanic) in Large Urban County and a Reference Group of White (non-Hispanic) mothers age 20-39, in 4 neighboring counties




Infection 4 5

Injury 12 3

Other 11 21


R95-SIDS 22 11



Grand Total 97 64


Are birth defects less common for Blacks, or

are they less diagnosed?

State Example CDC, NCHS, Linked Birth / Infant Death Records 2003-2006 Accessed at http://wonder.cdc.gov/lbd-current.html on Aug 19, 2011 1:02:49 PM State R95 (Sudden infant death syndrom - SIDS) 132 Other Q00-Q99 (Congenital malformations, deformations and chromosomal abnormalities) 55 Other V01-Y89 (External causes of morbidity and mortality) 51 R99 (Other ill-defined and unspecified causes of mortality) 24 Other & Suppressed 21

W75 (Accidental suffocation and strangulation in bed) 20 J00-J98 (Diseases of the respiratory system) 19

I00-I99 (Diseases of the circulatory system) 18 Q24.9 (Congenital malformation of heart, unspecified) 18 A00-B99 (Certain infectious and parasitic diseases) 16 G00-G98 (Diseases of the nervous system) 16 Other R00-R99 (Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified) 3

Denominator 273,899 (>1500 grams all Colorado residents)

Neighboring States Reference Group CDC, NCHS, Linked Birth / Infant Death Records 2003-2006 Accessed at http://wonder.cdc.gov/lbd-current.html on Aug 19, 2011 1:02:49 PM

Mountain States Reference Group

A00-B99 (Certain infectious and parasitic diseases) 13 G00-G98 (Diseases of the nervous system) 20 I00-I99 (Diseases of the circulatory system) 12 J00-J98 (Diseases of the respiratory system) 23 P00-P96 (Certain conditions originating in the perinatal period)

13

Other Q00-Q99 (Congenital malformations, deformations and chromosomal abnormalities)

59

Q24.9 (Congenital malformation of heart, unspecified) 20 Other R00-R99 (Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified)

7

R95 (Sudden infant death syndrom - SIDS) 112 R99 (Other ill-defined and unspecified causes of mortality) 25

Other V01-Y89 (External causes of morbidity and mortality) 44 W75 (Accidental suffocation and strangulation in bed) 12

All Other & suppressed 24 Denominator 397,082 (>1500 grams >12 yrs ed, >19 age, non-Hispanic White )

Compare States Reference Group Cause-Specific Mortality Rates (IH deaths per thousand live births)

CDC, NCHS, Linked Birth / Infant Death Records 2003-2006 Accessed at http://wonder.cdc.gov/lbd-current.html on Aug 19, 2011

State Neighbor States Reference Group

Suppressed, P00-P96, R00-R94 0.088 0.111 Infectious diseases 0.058 0.033 Nervous system diseases 0.058 0.050 Circulatory System diseases 0.066 0.030 Congenital malformations of the heart 0.066 0.050 Respiratory diseases 0.069 0.058 W75 Suffocation/Strangulation 0.073 0.030 R99 Other Ill-defined 0.088 0.063 V01-Y External causes 0.186 0.111 Q00-99 Other Birth Defects 0.201 0.149 R95 SIDS 0.482 0.282

TOTAL 1.435 0.856




2. “Risk and Preventive Factors”

● What are the primary risk and preventive factors KNOWN to be associated with the identified causal pathways or biological mechanisms?

● Which factors exhibits DISPARITIES in prevalence that reflect the observed outcome disparities?

Presenter

Presentation Notes

If the causal pathway is SUID, what are the primary reasons for excess SUID mortality? Are there differences in sleep position? Co-sleeping? Smoking? If the causal pathway is prematurity, does the target population have higher rates of chronic diseases? Higher rates of smoking? Higher infection rates? More stress? If the causal pathway is poor survival rates of vlbw babies, is the target population delivering high risk infants in hospitals without appropriate NICU facilities?

Phase 2 Analysis - Infant Health period Step 2: Estimate prevalence of risk and preventive

factors by type of mechanism

Infant Health

SIDS

Injury

Infection

Anomalies

Each cause category has its own set of risk and

preventive factors

Perinatal

Presenter

Presentation Notes

In terms of infant health, the first step is to investigate the underlying cause-of-death. Each cause-of-death category has its own specific set of risk or preventive factors. Therefore, further investigation can focus on specific underlying causes and their related risk and preventive factors. For SIDS, the factors might be the prevalence of infants sleeping on their belly or infants in a smoking environment. Again, this step investigates the reasons behind excess mortality so that interventions can be better targeted.

● Passive smoke ● Prematurity ● Sleep position ● Breast-feeding ● Bedding ● Co-sleep ● Maternal age ● Death scene

investigation

● Folic acid intake ● Alpha-feto protein ● Alcohol ● Drug abuse ● Diabetes ● Ultrasound ● Delivery site

SUID (& Sleep Injury?) Anomalies

Phase 2 Analysis-Infant health period Partial list of risk factors by major cause of death

Presenter

Presentation Notes

Each cause of death has probable risk and preventive factors. Even when SIDS is identified as the predominant cause, there are many risk factors that should be investigated.

● Medical home ● Prematurity ● Immunizations ● Breast-feeding ● Passive smoke ● Prenatal care ● Maternal age ● Infection type

● Bedding ● Carseats ● Supervision ● Environment ● Parent mental health/

substance use

Phase 2 Analysis-Infant health period Partial list of risk factors by major cause of death

Injury Infection

Presenter

Presentation Notes

The big advantage of going through the first step and identifying the predominant cause or causes is that the MANY risk factors for the other causes do not have to be investigated.

Phase 2 Analysis---Ficticiosia - Infant health period

● Step 2 examines risk and preventive factors for SUID and Injury

● Compares Study Population to Reference Population

● Population at risk is infants >1500 grams surviving 28 or more days

Disparities in prevalence of some risk factors for SIDS

from BRFSS (Urban County)

05

10152025303540

Inadequatesocial andemotionalsupport

Smoking Alcohol use(binge drinking)

BRFSS)

Low birthweight

BlackReference Group


● Tempting to pick the risk and preventive factors with the biggest disparity in prevalence

● Ideally should address factors with the biggest potential impact

● Estimating preventive impact of each factor on excess mortality helps prioritize the factors likely contributing to excess mortality

Presenter

Presentation Notes

Impact is a complicated issue. Some important factors are not modifiable, so we cannot hope to have an impact. Among factors that are theoretically modifiable, which have evidence-based interventions? How much can they be changed? What effect could that change have on the population ? (attributable risk) Which do we have local expertise about?

Infant Health Period Step 3: Estimate the impact of the risk and preventive factors

How much will the infant mortality rate in the study

population decrease if we decrease a risk factor Population Attributable Risk

Depends predominantly on: How “risky” the risk factor is?

Relative Risk

How many in the population are “exposed” to it? Prevalence

Population Attributable Risk Percent

• Compares rate for the whole population to the rate for those WITHOUT the risk factor

• Based on rate difference or (equivalently) on relative risk and prevalence of the exposure for the whole population.

• Interpretation: “Percent of the population that would be prevented from the poor outcome if the risk factor were eliminated from the entire population.”

• Relevant to estimating overall impact and cost.

Step 3: Estimate the impact of the risk and preventive factors

● Crude Estimate of Population Attributable Risk

● Population Attributable Risk Based on published adjusted relative risk

● Adjusted Population Attributable Risk using regression

● Which factors are modifiable? By how much?

Infant Health Period Step 3: Estimate the impact of the risk and preventive factors

“Disease” Not All Exposed a b n1 Unexposed c d n2

All a+c b+d n0

p2=c/n2 (rate of disease in low risk group)

p0=(a+c)/n0 (rate of disease in whole population) Levin’s PAF = (p0 - p2)/p0

Population Attributable Risk Percent

Step 3: Estimate the impact of the risk and preventive factors

Presenter

Presentation Notes

p0

15 sleep related deaths in a population of 8,550 were cross-tabulated by sleep safety and cause of death.

Cross-tabulate your data like this:

Infant Health Period--Hypothetical PAF Example

Sleep related death

Not sleep related death

Total

Exposed to risk (e.g. tummy-sleeping)

12 3408 5130

Unexposed (e.g. back sleeping)

3 5127 3420

Total 15 8535 8550

Presenter

Presentation Notes

p0

×

PAF answers the question: What if the whole population had the

lower, “safe sleep” rate of SUIDS deaths?

Infant Health Period--PAF example, continued

Mortality rate due to SIDS and unsafe sleep among those who slept safely

3 ÷ 3420 × 1,000

= .58 deaths per thousand

If all 8550 babies had slept safely, the estimated number of deaths is

.58 × 8,550

= 5 deaths

Compared to the actual 15 deaths, we would save 10, or 67% if we could make all babies sleep safely.

Presenter

Presentation Notes

p0

Step 3: Infant Health Period Estimate impact of the risk and preventive factors

Population Attributable Risk PAR = p0 – p2 = (p1 – p2)*n1/n0

PAF = (p0 – p2 )/p0 (x100 to get percent)

Alternate form – use if RR or OR is available PAF = P*(RR-1)/(1+P*(RR-1))

Limitation of PAF: In “real life”, factors do not act independently

Methods to deal with this issue, References from Deborah Rosenberg, PhD and Kristin Rankin, PhD

Division of Epidemiology and Biostatistics School of Public Health, University of Illinois at Chicago

Miettenin (1974) Adjusted PAF = Proportion of the disease that could be reduced by eliminating one risk factor, after controlling for others factors and accounting for effect modification

Bruzzi (1985)/Greenland and Drescher (1993) Summary PAF = Proportion of the disease that could be reduced by simultaneously eliminating multiple risk factors from the population . Method for using regression modeling to generate PAFs

Benichou and Gail (1990) Variance estimates for the adjusted and summary PAF based on the delta method

Eide and Gefeller (1995) Sequential PAF

Questions?

Phase 2 Analytic Methods Maternal Health/ Prematurity

Period of Risk

Presenter

Presentation Notes


0.8 1.1 0.6

2.0

EXCESS Fetal-Infant

Mortality Rate =4.4



MH/P 44%

MC 13%

NC 18%

IH 25%

Excess Mortality






Phase 2 Overview 52

MH/P period gap

Birthweight distribution

Maternal infection

Preconception health insurance

Maternal stress

Birthweight specific mortality

Referral to level III facility

Neonatal specialist availability

Phase 1

Phase 2 Step 1

Phase 2 Step 2

Maternal Health and Prematurity Risk Period




Presenter

Presentation Notes



Step 1. “Causal pathway”

● What causes of death contribute the most to excess mortality in this risk period?

● Can “patterns” in mortality disparities help us understand the underlying mechanism for excess mortality in this risk period?

Phase 2 Overview 55

Presenter

Presentation Notes

In the green box, which cause of death contributes the most to excess mortality? SUID? Infections? Accidents? In the blue box, are very low birthweight deaths due more to excess prematurity rates, or poorer survival rates?

Step 1: Identify Causal Pathways or Biologic Mechanisms for Excess Mortality

Cause of VLBW fetal and infant deaths is ● Multifactorial ● Complex ● Inconsistent ● Varies by training

ICD-10 Cause of Death Codes are not very

helpful

Maternal Health/Prematurity Period

Presenter

Presentation Notes

Analyses for Maternal Health/Prematurity approach this step differently. The underlying cause of death information for fetal and infant deaths born weighing less than 1,500 grams is not informative. Cause of death information in this weight range is complex and multi-factorial. The information is frequently inaccurate and inconsistently reported. Reporting varies by the perinatal capability of the hospital reporting and the clinical training of the certifier

A Tale of Two Cities

Nonicu City

Tinybaby City

1,000 births

1,000 births

10 VLBW deaths

10 VLBW deaths

For both cities, the "Blue Box" mortality rate is 10 deaths per thousand live births.

What can these cities do?

What is the difference between these two cities . . .

Let’s take a closer look

Nonicu City Tinybaby City 1,000 births 1,000 births 10 VLBW births 100 VLBW births 10 VLBW deaths 10 VLBW deaths

We were missing an important fact. The number of VLBW births sets these two cities apart.

What does this difference mean ?

Nonicu City

Tinybaby City

10 VLBW births 100 VLBW births 10 VLBW deaths 10 VLBW deaths Mortality rate for a baby born VLBW in Nonicu City is 100%

Mortality rate for a baby born VLBW in Tinybaby City is 10%

Kitagawa’s formula tells us: Which city we resemble, and What we need to focus on?

93%

7%

BirthweightDistribution

BirthweightSpecificMortality

Presenter

Presentation Notes

This partitioning is helpful because the factors and services that generally affect birthweight distribution are different from the factors and services that affect birthweight-specific mortality rates Should the community examine prevalence and impact of risk factors causing high VLBW/prematurity in their community, or should the community examine aspects of their perinatal care system that are responsible for higher birthweight-specific infant mortality rates in their community?

● Smoking ● Prenatal care ● Race ● Maternal age ● Parity ● Multiple Preg. ● SES/Education ● Birth Interval ● Maternal HTN/Diabetes ● Etc.

● Gestational age ● Referral system ● Perinatal care ● NICU system ● Mat. complications ● Neonatal conditions ● Pay source ● Etc.

Birthweight Distribution (VLBW Births)

Birthweight-Specific Mortality

Phase 2 Analysis—MH/P Period Partial list of risk factors by contributor

Presenter

Presentation Notes

These are some of the risk factors that should be considered, depending on which is the predominant causal pathway (based on Kitagawa analysis)

Kitagawa’s Formula Uses Algebra to PARTITION Excess Mortality into

An excel sheet at www.citymatch.org will do these calculations for you, if you give it some local data

∑

−×

++

−×

+nnn

nnnn

nn PPMMMMPP1 21

2121

21 )(2

)()(2

)(

1. Birthweight distribution

2. Birthweight specific mortality

Presenter

Presentation Notes

This partitioning is helpful because the factors and services that generally affect birthweight distribution are different from the factors and services that affect birthweight-specific mortality rates Should the community examine prevalence and impact of risk factors causing high VLBW/prematurity in their community, or should the community examine aspects of their perinatal care system that are responsible for higher birthweight-specific infant mortality rates in their community?

http://www.citymatch.org/

Phase 2 MHP period

Kitagawa Worksheet Data Entry Table 1: Target Population Enter Urban County Omaha Nebraska

Number of Number of Number of Live Infant Fetal Deaths

Birthweight Births Deaths 24+ wks 0-499 39 37 10 500-749 55 36 13 750-999 70 15 10 1,000-1,249 82 8 8 1,250-1,499 101 3 4 1,500-1,999 372 7 9 2,000-2,499 1,081 22 10 2,500+ 21,438 62 29

Urban Healthy Start PPOR Data 1997-1999

Urban Healthy Start Area

Reference

Opportunity Gap

10.7

2.1

8.6

3.8

2.7

3.6

.85

.85

.61

2.3

1.9

3.0

20.8 4.4 15.8 =

Urban Healthy Start Area 1997-1999

Birthweight Distribution

73%

Birthweight Specific Mortality

27%

Kitagawa Partitioning of Excess Mortality in the MH/P Period of Risk

Kitagawa Analysis (Birthweight under 1500 grams) African Americans in Example City vs

U.S. Reference Group

92.5%

7.5%

BirthweightDistribution

BirthweightSpecificMortality

Maternal Health/Prematurity Period

Presenter

Presentation Notes

Here we see a typical city, where 91% of the VLBW excess mortality is due to BIRTHWEIGHT DISTRIBUTION. The predominant underlying cause of death for VLBW babies is birthweight distribution – too many babies are born at very low weights. Our community will benefit most by preventing prematurity Some cities find that a significant proportion of their VLBW excess mortality is due to birthweight specific mortality. These cities must explore risk factors related to perinatal systems. Birthweight specific mortality nearly matches that of the reference group, so babies that are born too small are surviving as well as can be expected in our community.



● Smoking ● Prenatal care ● Race ● Maternal age ● Parity ● Multiple Preg. ● SES/Education ● Birth Interval ● Maternal HTN/Diabetes ● Etc.

● Gestational age ● Referral system ● Perinatal care ● NICU system ● Mat. complications ● Neonatal conditions ● Pay source ● Etc.

Birthweight Distribution (VLBW Births)

Birthweight-Specific Mortality

Example City can focus on causes for “TOO MANY VLBW BIRTHS”

X

Presenter

Presentation Notes

These are some of the risk factors that should be considered, depending on which is the predominant causal pathway (based on Kitagawa analysis)

Maternal Health/Prematurity Step 2: Analysis plan depends on result of Kitagawa

Maternal Health/ Prematurity

Birthweight Distribution

Birthweight- Specific Mortality

Mortality Rate

OUTCOME POPULATION AT RISK

Percent VLBW All Births And Fetal

Deaths

VLBW Births and Fetal Deaths

Presenter

Presentation Notes

If the predominant component is birthweight distribution, the focus of the PPOR analysis shifts from studying the deaths of these small babies to studying the percentage of all births weighing 500 to 1,499 grams or very low birthweight live births (VLBW). The analysis examines the prevalence of risk and preventative factors for VLBW births among all live births, which is often available from sources other than vital records, allowing analysis of additional factors.





Phase 2 Overview 71

Questions?

Phase 2 Analytic Methods Maternal Health/ Prematurity

Period of Risk

Presenter

Presentation Notes


0.8 1.1 0.6

2.0

EXCESS Fetal-Infant

Mortality Rate =4.4



MH/P 44%

MC 13%

NC 18%

IH 25%

Excess Mortality





Presenter

Presentation Notes


Maternal Care Period Phase 2 Analytic Strategy

● Less needed information on fetal death certificates ● More data is missing/not reported on certificate ● Missing major causal pathways on certificate

Chromosomal abnormalities Severe congenital anomalies Placental vascular abnormalities Antibodies

● Need to adjust PPOR Phase 2 analytic strategy and steps

Phase 2 Analysis 76

Presenter

Presentation Notes

Many state vital records systems do not collect as much information on fetal deaths as on live births, have a high frequency of missing information on fetal deaths, or do not conduct as much editing or querying of information collected on fetal deaths. Defining a causal pathway is more difficult than for the Infant Health period because the current cause of death information for fetal deaths is not as useful in developing prevention strategies





Phase 2 Overview 77

XXX

Maternal Care Period Step 2: Prevalence Comparison of Risk and Preventive Factors

Potential known reliable risk & preventive factors: ● Maternal age and race ● Education and socioeconomic ● Parity and previous fetal loss ● Inter-pregnancy interval ● Smoking ● Prenatal care ● Multiple gestation ● Birthweight and Gestational age ● Medical Conditions (Diabetes, hypertension, RH disease) ● BMI and gestational weight gain ● Genetic testing

Presenter

Presentation Notes

A few known risk factors t for death in the Maternal Care period, thought to be reliably collected in some states, are birthweight, gestational age, maternal age, race, parity, smoking, education socio-economic factors, previous fetal loss, interpregnancy interval and multiple-gestation pregnancies. Population-based risk factors from other health information systems include body-mass index, weight gain during pregnancy adjusted by body-mass index, and other select maternal medical conditions (such as, diabetes, hypertension, RH disease, etc.). Some of these risk factors may be collected on the newly revised fetal death certificate, WIC program database, and other public health datasets. With currently available data, the best investigative approach for excess Maternal Care deaths may be an examination of the known risk factors using vital records and other data sources, and conducting fetal mortality reviews on these larger fetal deaths from a health prevention perspective.

0

20

40

60

80

100

120

Currentsmoker

Heavydrinking

Report goodhealth

HealthInsurance

HS/GED

Black N.H.

Reference Group

Phase 2 MHP period

Disparities in prevalence of risk factors (Urban County BRFSS— represents all females age 18-44)

0

10

20

30

40

50

60

70

Black N.H.

Reference Group

Phase 2 MHP period

Disparities in prevalence of risk factors, (Urban County vital records data – PPOR eligible live births plus fetal deaths)

Presenter

Presentation Notes

Matern Child Health J. 2010 Jan 16. [Epub ahead of print] Assessing Maternal Risk for Fetal-Infant Mortality: A Population-Based Study to Prioritize Risk Reduction in a Healthy Start Community. Kothari CL, Wendt A, Liggins O, Overton J, Del Carmen Sweezy L. Research Department & Emergency Department, Kalamazoo Center for Medical Studies, Michigan State University, Kalamazoo, MI, USA, [email protected]. Abstract Study goals were to distinguish between maternal risk factors for fetal versus infant mortality, and to identify which maternal characteristics contributed the greatest risk of mortality overall. This case-control retrospective study abstracted data on more than forty maternal characteristics from 261 prenatal and delivery records: all 26 fetal deaths, all 40 infant deaths and 195 randomly selected surviving births in a high-mortality Healthy Start community. Bivariate and multivariate analyses were conducted. The fetal-mortality population was significantly more likely than the infant-mortality population to have no insurance (P = .047), inadequate prenatal care (P = .039) and previous fetal death (P = .021). Comparing the combined mortality population with the surviving sample, two tiers of risk emerged: Rare-but-lethal risks, including no prenatal care (P < .001) and Child-Protective-Service involvement (P = .001), and common-and-dangerous risks, including inadequate maternal weight gain (OR = 13.55), drug or alcohol abuse (OR = 8.67), obesity (OR = 2.77) and anemia (OR = 3.61). Both fetal and infant mortality groups must be considered when identifying maternal risks. Inadequate prenatal weight gain, obesity and anemia contribute as much to feto-infant mortality as substance abuse. Public health efforts to improve maternal nutrition and healthy weight should be redoubled. PMID: 20082128 [PubMed - as supplied by publisher] LinkOut - more resources





Phase 2 Overview 81

XXX

Questions?

Documents

Perinatal Periods of Risk Practice Collaborative · Phase 1 results suggest some general areas for action, but are not sufficient information alone. \爀䘀漀爀 䴀愀琀攀爀渀愀氀

Perinatal Periods of Risk Practice Collaborative · Phase 1 results suggest some general areas for action, but are not sufficient information alone. \爀䘀漀爀䴀愀琀攀爀渀愀氀