Data Sets

Abortion

• Does the Abortion Rate include Fetal Deaths?
  • NO, abortion rate, pregnancy rate and abortion percent DO NOT include fetal deaths.   If you are looking at CHS tables note that while abortion tables DO NOT include fetal deaths, pregnancy tables DO.  Therefore, VistaPHw numbers may not match CHS numbers.

Birth

• Why are the calculated gestational age data different from previous versions of VistaPHw?

• The gestational age in weeks is calculated by subtracting the date of last normal menses from the birth date, dividing by 7 and truncating the result to eliminate decimal places.  If the menses day is missing but the month and year are present, a value of '15' is used for the day.

In previous data files (issued before 2006), if the menses month and/or year were missing or the calculated gestational age was beyond a reasonable range (<18 or >45 weeks), the gestational age was estimated from the child's birth weight.  

Currently, if the gestational age cannot be calculated because of missing menses dates or the calculated age is out of range, the clinical estimate is used.  If the clinical estimate is also out of range or unknown, the calculated age is unknown.  For 1980-88, the birth certificate did not collect the clinical estimate.  For these years, only the gestational age calculated from the menses date is included in this field.  In all other cases, the calculated age is unknown.

This change makes the Washington State data consistent with data published by the National Center for Health Statistics (NCHS).  

• Why is maternal smoking by trimester and BMI only available after 2003?
  • Maternal smoking by trimester and BMI were added to the 2003 revised birth certificate. We do not have data prior to that time.

   

• Why are anemia and alcohol use no longer available?
  • Maternal anemia and alcohol use are not collected in the 2003 revised birth certificate.

CHARS

• Why doesn't CHARS data include RACE?

  All of the hospitalization data comes from billing data.  It is not required to report a race on billing forms when a hospitalization occurs.

• What is the difference between CHARS duplicated vs. unduplicated data?

  VistaPHw currently includes hospitalization data (CHARS data) which consists of all discharges from WA hospitals in a given year.  Some patients may be hospitalized and discharged more than once in any given year so they appear more than once in the data. In the past, an unduplicated dataset has been available, which allows the data to be analyzed by patient rather than hospitalization. This file is under review by the DOH Center for Health Statistics.

• Does the hospitalization data in VistaPHw include residents outside of Washington?

  No, the hospitalization data in VistaPHw only includes Washington state residents.

  We are currently exploring the addition of hospitalizations of WA residents in Oregon, Veteran Administration, and military hospitals.   

Death

• How are the underlying causes of death grouped in VistaPHw?

  Due to the detailed nature of the ICD classification system, it is common to group ICD codes into more general categories for analysis and comparison purposes. VistaPHw uses the classification system provided by the National Center for Health Statistics (NCHS) in which causes of deaths for adults are grouped into 113 separate groups and deaths for infants into 130 groups. NCHS groupings are also used in the Washington State Center for Health Statistics publications.

Sub-County Population

·         The Data Set is an "Unofficial Estimate", what does that mean?

The unofficial sub-county estimates were developed for public health assessment only, and not for estimating the population of cities, towns, and counties.  Official population estimates for cities, towns, and counties are produced under the authority of the Office of Financial Management (OFM) in accordance with RCW 43.62.020 and RCW 36.12.100.  The OFM estimates are developed for official state planning, funding and program administration issues. Under no circumstances should the "unofficial" sub-county population estimates be used for these purposes.

·           What levels of geography are available in VistaPHw?

Each Local Health Jurisdiction has access to sub-county estimates specific to their geographic region.  Currently, all LHJs have access to their own zip code, census tract, and block group denominator estimates and hospitalization numerator data.  This can be accessed by changing the Data Collection Folder from WA to your LHJ. Please contact the VistaPHw coordinator if you are not able to access your local data.

·         The estimates produce differing rates, why is that? OR I ran a new year of data in VistaPHw and there is a large difference between the old data and the new year of data. What's happened?

For each iteration of the population estimates, the new estimates in VistaPHw supersede the old population estimates, so that the same methodology of estimation is applied to all years of data. We recommend re-running historical data for comparison. Since the entire denominator data (population) changed, the rates will be slightly different. So while we expect you to see some differences in the rates, if you notice something unusual, please contact the VistaPHw Coordinator at Buffi.LaDue@doh.wa.gov .

·         OFM does population estimation,  so why is there a different set?

Subcounty population estimates produced by OFM are population totals for incorporated cities and towns for allocation of state revenues and program administration. OFM produces age by sex estimates and race estimates independent of each other but only at the county level- that is, they do not produce sex by age by race estimates at the subcounty level. The subcounty estimates here are adjusted to county race and sex by age totals provided by OFM but are redistributed according to census counts at the block and blockgroup levels, and include housing starts, to produce the population age by sex by race components routinely used in subcounty public health data analysis and community health assessment.

Top

Epidemiology

• Why should I use 2000 vs. 1940 Age-adjusted standards?
  • Age-adjusted rates are used when you want to do comparisons.  For example, you may want to compare your rate to another county's rate or to the state rate where the age structures of the two populations may be different.  To compare the rates between these areas, you need to use the same standard population. Currently, the most commonly used standard is the 2000 US Standard population.   However, you should always check to make sure the age-adjusted rates you are using for the comparison is based on the same standard.  (Some cancer data is age adjusted to the 1970 US population.)   If you use the 1940 US population as your standard but another county uses the 2000 US population, then these age-adjusted rates cannot be compared.

Because the population has changed since 1940 it makes sense to update the standard population that everyone should use in order to draw comparisons.  In 1999, the US Standard population changed to a projected year 2000 age distribution.   This population distribution is closer to the actual age structures of most areas in the late 1990s compared to the 1940 age distribution which has fewer elderly people.  As a result, age-adjusted rates using the year 2000 standard may appear closer to the crude rates.  It is important to remember that age-adjusted rates are still hypothetical rates.  Specifically, an age-adjusted rate is the one that would occur if your county had the same age structure as the standard population.

• What is the difference between generating counts using a crude rate or an age-adjusted rate?
  • If you want to get a number of events in a particular population, or count, it is appropriate to use the crude rate statistic.  This will give you the total number of events.  If you use the age-adjusted rate, the count will exclude any event where age was not provided.  So, the total number of events will probably be lower than the actual count.

   

• Why are the counts for the state total different for Age-adjusted vs. Crude rates?
  • An age-adjusted rate is calculated by computing a weighted average of age-specific rates.  On the other hand, crude rates do not include age.   Cases where age is unknown can be included in the numerator for crude rates but, not age-adjusted rates.  This makes the total count higher for the crude rate.   The age-specific and age-adjusted counts should be equal.

   

• Can differing Population Denominator sources cause discrepancies in rates?
  • Population denominators can be the probable cause of the discrepancy in rates. We have found that even a small difference in denominators, such as using the original OFM series in rounded or unrounded format, is enough to produce such discrepancies in rates.

• Why do population counts for the same geographic area sometimes seem to differ?
  • Counties are defined two ways. One uses Census boundaries (“actual”) versus a ZIP-defined county boundary. ZIP code areas often cross county boundaries. This means a ZIP code definition of a county can differ from the actual boundary, and the population counts differ because they are defined by different geographic boundaries. This is most frequently seen in hospitalization data. In this assessment topic,  the population figures are given for ZIP code boundaries--and may be aggregated to the group of ZIP codes that most closely define a county's boundary--because hospitalization numerator data (i.e., counts of hospitalizations) are only available by ZIP code.

   

• What is the difference between BINOMIAL and POISSON Confidence Intervals?
  • The 1994 PHIP recommends that if there are 100 or fewer events or the rate is less than 10%, 95% Poisson confidence intervals should be calculated. If there are more than 100 events and the rate is greater than 10%, then the normal approximation to the binomial is recommended (in VistaPHw this is labeled "binomial").
    
    
  • In reality, the appropriate distribution that describes these outcomes for people (i.e., death/no death, birth/no birth, low birthweight/normal birthweight, etc.) is the binomial distribution. Calculation of exact binomial confidence intervals involves complex mathematics. However, conveniently, for rare events, the Poisson distribution (by Ury & Wiggins) approximates the binomial distribution, and for non-rare events the normal distribution approximates the binomial distribution. It is relatively straightforward to calculate confidence intervals for Poisson and normal variables, so these methods are in VistaPHw. Thus, in VistaPHw, when "binomial" confidence intervals are specified, what is actually calculated are confidence intervals based on the normal approximation to the binomial distribution rather than exact binomial confidence intervals. 
    
    As for the PHIP recommendations, it is necessary to have a universal rule of thumb, although, like most of these general recommendations, it's probably not always correct; for instance, in runs recently done with King County data it seemed to make more sense to start using binomial CIs at a 5% prevalence (calculated with Stata 5.0 cii command).  But the results for every county would probably be different.  If alot is riding on where the CIs fall and the CIs are very close to overlapping, it's probably a good idea to calculate exact binomial CIs. In practice it probably won't make a difference because the difference between the two methods usually is, at most, in the first decimal place of the rate.

Another minor note about terminology: the terms confidence intervals and confidence limits are different. VistaPHw calculates both confidence limits and the confidence interval.   The confidence limits are labeled "UB" for Upper Bound and "LB" for Lower Bound in the excel output.  The confidence interval is the distance between the upper and lower bound of the confidence limits.   The "upper interval" is the distance from the rate to the upper bound (analogous definition for "lower interval"). 

Top

Error Messages

• Error 339

Problem is most likely a 'rights' problem - an administrator needs to allow the user a higher level of access

• Error 380

Problem can be fixed by uninstalling/reinstalling or deleting the C:\Program Files\VISTAPHW\USA folder and all its subfolders, then logging back on to VistaPHw

• Error 1391

The windows installer service cannot update the file c:/winin/system32/scrrun.dll because the file is protected by windows.  You may need to update your OS.

Many people in the health assessment field have had this problem because it is related to EpiInfo 2000.

Solution:

After both systems are installed, you can repair the EpiInfo mischief by re-registering the system version of scrrun.dll.  Both systems should then function properly.  Here's how: 

1.  click Start, Run.

2.  type  "Regsvr32 c:\winnt\system32\scrrun.dll" (without quotes).

3.  press Enter, wait for success message.

 If you have a system other than Windows 2000 and encounter this problem, use the same solution except in step 2 change the path to the location of scrrun.dll on your hard drive.

Top

---

ICD-10 Codes

• What is the significance of the implementation of ICD-10 code?

Beginning with deaths occurring in January 1999, the United States began using International Classification of Diseases (ICD-10) to classify causes of death reported on death certificates.  Prior to this, ICD-9 was used from1979-1998.  Implementation of ICD-10 has had an important impact on the presentation and interpretation of mortality statistics by cause-of-death.  The change to ICD-10 created a discontinuity in trends that must be accounted for when comparing mortality during 1999 and later to prior years.  To put it another way, cause-of-death data for 1999 and later years are not comparable to prior years, unless adjustments are made for the coding and classification changes.  Without adjustment, it is impossible to know whether an observed increase or decrease in deaths due to a particular cause is "real" or merely the result of the changes in classification and coding.  For further information on the changes, please click on the links below:

Washington State, Center for Health Statistics, ICD-10 Information Page

National Center for Health Statistics web site: http://www.cdc.gov/nchs/about/major/dvs/icd10des.htm

• How do you adjust for the change between ICD-9 and ICD-10?

To enable comparisons across the ICD-9 to ICD-10 transition, a comparability study was carried out by the National Center for Health Statistics (NCHS). NCHS double-coded large samples of the 1996 national mortality file, once by ICD-9, and again by ICD-10. A comparability ratio was then calculated by dividing the number of deaths for a selected cause of death classified by ICD-10 by the number of deaths classified to the most

nearly comparable cause of death by ICD-9. The resulting ratio can be used to adjust counts and rates for a given cause of death classified by ICD-9 so they are comparable to those for the most similar cause classified by ICD-10. The ratio will also allow users to estimate the extent of the discontinuity of the change to ICD-10 by showing the

net effect of coding and classification changes.