Rachel,

Test/Validation 1: Assurance that the system can correctly generate and
interpret the base set of X12 rules. Without this assurance everything else
is a house of cards. The system must then be able to identify and report
back to the originator errors detected with sufficient detail such that the
originator can identify and resolve the problem.

#1 - Yes I agree with this test scenario although its implementation is not
cut and dry. Several issues arise here, the validation software being used
today gives different results, how will you ever know you are completely
correct? Secondly, there is a large test data need in order to fully and
robustly test all the various conditions. If you are creating software then
you must test it for all possible variations and implementations, hence the
need for a structured decomposition of the software and subsequent test case
and test file creation process.
Test/Validation 2: Assurance that the system can now generate and interpret
X12 interchanges containing functional groups and transactions that both
comply with the X12 rules AND the specifications of the related HIPAA
implementation guides. The system must be able to identify and report back
to the originator errors detected that render the interchange and included
functional groups and transactions NON-COMPLIANT with the HIPAA
specifications. This is also essential since there are economic penalties
set forth in the law for conducting non-standard transactions.

#2 - All the above issues and the creation of structured negative test
conditions specifically designed to cause front-end rejection by the
validation software and the translators.
Test/Validation 3: Assurance that the system can generate and interpret X12
interchanges containing functional groups and transactions that now contain
the sufficiently correct standard data that will ensure the receiving party
can successfully process the transactions through its production systems to
the desired outcome: payment of a claim, an accurate response to an inquiry
re eligibility or claim status or referral authorization.

#3 - This requires massive amounts of production test data to prove business
process compliance. Stakeholders will be reviewing internal test results to
determine compliance and whether or not their applications are functioning
properly. Of the three examples, the only one that has a baseline is the
production system. Even though a production system may contain defects, they
are known entities and can be accounted for in the expected results and
concurred by the business. Validation and translators have unknown errors
and in their current state some of them cannot be used as a baseline.
So, then, what do you recommend the industry do to achieve these three
fundamental assurances - that is, assuming you agree with them?

Options-
1)      Prior to contracting with a validation service maybe clients should
review internal reports on the testing effort of validation software and see
if the defects discovered have been corrected and review the robustness of
the test effort in general. If validation companies are preaching testing
methodologies and come across as testing experts then shouldn't they have
taken what they preach and used it internally first? Wouldn't you think they
would have had to use an extremely large test file repository to test their
own software, since it must accurately report on every possible defect in
X12 and the IG's? The answers may surprise you.....  I personally use
validators to provide feedback for remediation but under no condition will I
treat what they say as gospel. Validation software cannot certify anything
until the software that runs it has been certified as functioning properly.
2)      In regards to translators, the above statement should be true but to a
lesser extent, since client mapping is involved and they cannot truly be
completely tested on behalf of the covered entity until the mapping has been
completed. Although, a translator can be tested for an off-the-shelf
compliance for X12 and IG's. What an end user does with the product after it
has been installed is a whole different story.
3)      For software vendors, they can also create an off-the-shelf compliance by
thoroughly testing their products prior to release. If the software is just
installed without modifications then the compliance still holds, if changes
are made then testing from a covered entity perspective must be performed
on-site.
4)      A topic not talked about enough is the level of testing that
clearinghouses have to perform, maybe there should be a whitepaper on that
topic alone.
5)      Since I believe in a free market system, I don't believe WEDI should do
anything on behalf of free test files. Who volunteered WEDI to do that in
the first place? Myself and other companies happen to sell test files and we
strictly prohibit their use outside of an entity. Why should WEDI give
something for free? They don't give free translators or free EDI validation
products for public consumption, the same should hold true for any
marketable product or service. If individual companies want to go through
all the effort to catalog, document and give away free files then let them
do it on their own and deal with any circumstances arising from their
decision, but do not expose WEDI to any adverse situations that may arise.
6)      HIPAA is unique in one regard, organizations cannot just release software
with defects and fix the rest in production like they do with most any other
release. If these entities are using old paradigms for testing, then it is
time for them to realize that HIPAA does require an extremely structured and
robust testing effort. Concepts like test automation and test management
tools should be considered since HIPAA remediation will continue for quite
some time and they will need reusability for all their testing efforts.  I
don't foresee the phrase, "we are 80% compliant or 90% compliant", and you
either are or are not.
7)      I also think there should much more involvement in this forum from
quality assurance and testing professionals, resources that have tested
large legacy and client server applications and know the amount of work
involved to prove a product is in compliance or certifiable. I have made a
concerted push to my fellow QA professionals for greater involvement and
gratefully I can say that they will become more prominent within the next
few weeks.
8)      I don't feel I am in the minority and by the number of people I have
discussed this issue with, they all agree something must be done, done
quickly and done right. The solutions to these issues lie within the
industry itself and its ability to self regulate, there are some great ideas
being circulated and I hope to see them realized very shortly.

Mark A Lott

 -----Original Message-----
From:   Rachel Foerster [mailto:[EMAIL PROTECTED]]
Sent:   Sunday, August 11, 2002 9:55 AM
To:     [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject:        RE: Test Data

Ok, Mark. You've set forth the philosophy for rigorous, controlled,
disciplined testing. Given the health care industry, size, fragmentation,
etc. what do you believe should be the approach? I've said before, and I
continue to say it that there are several levels of testing and validation
that need to take place - all under controlled test plans and scenarios. The
first one is essential for starters since I believe there are many, many
small vendors developing software for HIPAA covered entities to use that are
incorporating rules for generating and interpreting the X12 rules on which
the HIPAA specifications are based. So,

Test/Validation 1: Assurance that the system can correctly generate and
interpret the base set of X12 rules. Without this assurance everything else
is a house of cards. The system must then be able to identify and report
back to the originator errors detected with sufficient detail such that the
originator can identify and resolve the problem.

Test/Validation 2: Assurance that the system can now generate and interpret
X12 interchanges containing functional groups and transactions that both
comply with the X12 rules AND the specifications of the related HIPAA
implementation guides. The system must be able to identify and report back
to the originator errors detected that render the interchange and included
functional groups and transactions NON-COMPLIANT with the HIPAA
specifications. This is also essential since there are economic penalties
set forth in the law for conducting non-standard transactions.

Test/Validation 3: Assurance that the system can generate and interpret X12
interchanges containing functional groups and transactions that now contain
the sufficiently correct standard data that will ensure the receiving party
can successfully process the transactions through its production systems to
the desired outcome: payment of a claim, an accurate response to an inquiry
re eligibility or claim status or referral authorization.

So, then, what do you recommend the industry do to achieve these three
fundamental assurances - that is, assuming you agree with them?

Rachel
Rachel Foerster
Principal
Rachel Foerster & Associates, Ltd.
Professionals in EDI & Electronic Commerce
39432 North Avenue
Beach Park, IL 60099
Phone: 847-872-8070
Fax: 847-872-6860
http://www.rfa-edi.com


-----Original Message-----
From: Mark A Lott [mailto:[EMAIL PROTECTED]]
Sent: Saturday, August 10, 2002 6:28 PM
To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: RE: Test Data


Greetings,

It seems the debate over test files gets livelier everyday, quite refreshing
I think. To me it shows the difference between initial concept and actual
implementation of such an initiative. I want apologize up front for the
length of this e-mail but this subject is a complex one to explain in a
brief synopsis. The views expressed here are my own and come from
considerable time involved in the Quality Assurance world, with heavy
emphasis on EDI , financial transactions and certifying millions of file
formats and trading partners across the globe. I am a testing purist and
feel too many organizations really don't understand the complexities and
work involved in the HIPAA testing arena. Some reasons for this are from
misinformation being published, trying to simplify a very time-consuming and
difficult task of testing, and the belief that EDI validation is somehow the
center of the testing universe even though it represents just a very small
piece of the testing lifecycle. There is no such entity as an EDI Validation
Test Phase, yet we choose to separate it from the overall testing process.
EDI validation occurs at every level of testing and the test files really
need to reflect the purpose for which it is being used in conjunction with
the appropriate test phase. I for one, do not agree that free test files are
needed and considering the enormous time commitment required to properly
catalog and describe in detail what each file represents, it realistically
cannot be accomplished in a meaningful way for the industry. This comes with
one caveat, free test files would be beneficial for unit testing and file
formatting in the developer arena but not really a viable alternative for a
test team trying to conduct Integration, User Acceptance or Trading Partner
testing.

It begs the question; why do people feel test files should be free and what
purpose are you trying to fulfill? Free test files will in the end cost more
money than if you started from scratch. Personally, I believe that all test
files should be created using production data especially for traceability
and to benefit business stakeholders during the review of business
functionality during the user acceptance and trading partner test phases.
Creating test files using an EDI tool or via your internal translator
introduces additional errors into the test verification process that will
only cause expanded testing delays, timelines being missed and a
misconception that your systems have been fully tested. I have also heard
phrases like "my translator or software is HIPAA compliant, so we are okay".
All I can do is wonder why people believe all these misconceptions. "If you
tell a lie long enough, it will still never be considered a fact"

Allow me to digress on this subject for people without formal QA training
and expertise. A test file is derived from a test case which is derived from
a set of test conditions which is created in response to testing and proving
a business requirement. The test case must have an expected result in order
to prove proper functionality. Free test files do not have expected results
and in essence would have to be retrofitted into the test conditions. Using
a free file causes test analysts to decompose the file to determine all of
the pertinent contents, then adjusting the test file to meet test cases and
conditions. A very important point is that test files, test cases and test
conditions must have traceability to be usable, otherwise the test team has
to look everywhere for defect correction when a discrepancy is found. Is the
defect in the translator, the EDI validation tool, in the test file itself,
in the understanding of the guides, was the business requirement written
correctly, etc...? In reality, available free test files are only an
infinitesimal representation of all the conditions that need to be tested to
correctly ascertain whether or not business functionality has been proven. I
am not sure there is great benefit in providing a test analyst 1% - 5% of
their overall test file needs. There was an estimate of 1400 possible test
files for Oxygen Therapy (both positive and negative test conditions), care
to guess how many organizations executed even one quarter of the possible
files? QA personnel should know better than to assume any particular aspect
or product is 100% correct.

Now, let's look at the bigger picture which will cause even more contention.
If you offer free test files, through which validation engine where they
executed? EDI validation tools all give different answers, but which one is
right? All of them? None of them? A combination of all of them? Lets examine
a real scenario: Payer X using EDI tool A to verify their files,
Clearinghouse Y uses EDI tool B to verify their files, Providers X,Y and Z
all use different EDI tools for their validation. All covered entities
believe they are compliant because their EDI tool told them so. Provider Y
sends what they believe to be a complaint file to Payer X, yet the
transaction is rejected by Payer X because their front end EDI tool says it
is not compliant, and so on and so forth...  Can you see what the industry
will uncover, but by that time it is too late.

There is also discussion about having all the EDI vendors perform testing
with each other, why? Shouldn't they have tested their product completely
before they released it to the industry? If a company produces a quality
product they should be rewarded for doing so, and quality usually comes with
a price. Last I checked, this is still a free market society. This should
also prove the point that EDI validation is just a small part of the
information chain and cannot be seen as a so-called verification or
certification point. The same goes for translators as well. These are all
various pieces making up a whole and I highly doubt anyone of them are
defect free. Covered Entities MUST examine testing from a business process
and business functionality perspective.

So many of us with years in quality assurance agree that you cannot stress
enough that you must start from the known and progress to the unknown, not
from unknown to unknown and NEVER assume any piece of software is 100%
defect free. Therefore you should have test conditions that represent
application test data, specific translator test data, EDI validation test
data, etc. If you start with trying to prove business functionality in
conjunction with a very robust test effort, you will uncover translator and
EDI validation defects. If you start from EDI validation and go backwards,
you will not find all your defects and will leave your organization exposed.
How does an organization know that their systems have been fully tested?
Certainly not by running free test files through their application systems,
although there may very well be people doing this. By offering free files
you open yourself up to many, many issues concerning viability, traceability
and responsibility. Is this really the best use of volunteer time? I don't
believe it is. I fear that true testing methodologies, concepts and best
practices are not being promulgated enough in this industry and in this
forum and organizations will have a false sense of security about their
applications systems.

I strongly believe that people need to really open their eyes and see what
is happening in this industry and not assume everything will be taken care
of for them. There are many organizations looking for help to comply with
HIPAA transactions and code sets, and they look to us for assistance, we
must try to not get caught up with our own business initiatives and do what
it right for the industry as a whole, isn't that the definition of
volunteerism, to help the greater good? This in no way reflects on the all
the volunteers that have graciously given their time to work for the
betterment of the industry, but a much more realistic and hard-hitting
approach needs to be taken with a focus on true implementation if we are to
expect this venture to be a successful one. I apologize for leaving off the
sugar coating, but their needs to be a line drawn in the sand not a hole to
bury one's head. This only reflects the tip of the iceberg and the
frustration currently being voiced on the current testing implementation
shortcomings will only grow and are very much welcome to assist in driving
the point home. New education and industry initiatives are needed...

If I may share a few words of wisdom:

"Failure is not a single, cataclysmic event.
    You don't fail overnight. Instead, failure
    is a few errors in judgment, repeated every
    day." -- Jim Rohn

"The quality of the solution you pick will be in
    direct proportion to the quantity of the
    solutions you consider." -- Brian Tracy


Mark A Lott


Please note that it may take up to 72 hours to process your request.
<P>The WEDI SNIP listserv to which you are subscribed is not moderated.  The
discussions on this listserv therefore represent the views of the individual
participants, and do not necessarily represent the views of the WEDI Board
of
Directors nor WEDI SNIP.  If you wish to receive an official opinion, post
your question to the WEDI SNIP Issues Database at
http://snip.wedi.org/tracking/.
Posting of advertisements or other commercial use of this listserv is
specifically prohibited.



Please note that it may take up to 72 hours to process your request.
<P>The WEDI SNIP listserv to which you are subscribed is not moderated.  The
discussions on this listserv therefore represent the views of the individual
participants, and do not necessarily represent the views of the WEDI Board of
Directors nor WEDI SNIP.  If you wish to receive an official opinion, post
your question to the WEDI SNIP Issues Database at
http://snip.wedi.org/tracking/.
Posting of advertisements or other commercial use of this listserv is
specifically prohibited.

Reply via email to