Blood Safety: Enhancing Safeguards Would Strengthen the Nation's Blood
Supply (Testimony, 06/05/97, GAO/T-HEHS-97-143).

GAO discussed its two reports on the safety of the nation's blood
supply, focusing on: (1) the current risks of blood transfusion; (2) the
content and quality of data collected to assess these risks; and (3) the
Food and Drug Administration's (FDA) layers of safety and their ability
to ensure the safety of the nation's blood supply.

GAO noted that: (1) its analysis of current risks from transfusion
showed that, while the nation's blood supply is safer today than at any
time in recent history, some risk remains, even if all the safeguards
available work perfectly; (2) GAO also found several vulnerabilities and
gaps in current procedures which, if eliminated, would provide greater
assurance of safety for the nation's blood supply; (3) the most serious
of these problems follow: (a) not all donors who test positive for
certain viruses are notified, which means that they can attempt to
donate again and also may go without treatment; (b) similarly, not all
recipients of virally contaminated blood are notified, which may keep
them from seeking treatment and also allow them to transmit the disease;
(c) blood facilities are not required to remove from their inventory
blood from donors who have subsequently tested positive for viral
infections; (d) unlicensed blood facilities that, together, produce 10
percent of the nation's blood do not have to submit to FDA error and
accident reports that may signal the need to recall potentially
contaminated units of blood; (e) FDA's investigations of error and
accident reports that warrant a recall take a long time and increase the
risk that units will have been transfused before a recall is
accomplished; and (f) finally, FDA's inspections of blood facilities are
inconsistent in focus, scope, and documentation; and (4) GAO's reports
contained a number of recommendations to the Secretary of Health and
Human Services to eliminate these weaknesses in the quality assurance
system for the blood supply.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  T-HEHS-97-143
     TITLE:  Blood Safety: Enhancing Safeguards Would Strengthen the 
             Nation's Blood Supply
      DATE:  06/05/97
   SUBJECT:  Product safety
             Infectious diseases
             Quality control
             Testing
             Acquired immunodeficiency syndrome
             Safety regulation
             Health hazards
             Inspection
             Reporting requirements
             Medical supplies
IDENTIFIER:  AIDS
             FDA Error and Accident Reporting System
             
**************************************************************************
* This file contains an ASCII representation of the text of a GAO        *
* report.  Delineations within the text indicating chapter titles,       *
* headings, and bullets are preserved.  Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are           *
* identified by double and single lines.  The numbers on the right end   *
* of these lines indicate the position of each of the subsections in the *
* document outline.  These numbers do NOT correspond with the page       *
* numbers of the printed product.                                        *
*                                                                        *
* No attempt has been made to display graphic images, although figure    *
* captions are reproduced. Tables are included, but may not resemble     *
* those in the printed version.                                          *
*                                                                        *
* A printed copy of this report may be obtained from the GAO Document    *
* Distribution Facility by calling (202) 512-6000, by faxing your        *
* request to (301) 258-4066, or by writing to P.O. Box 6015,             *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time.                                    *
**************************************************************************


Cover
================================================================ COVER


Before the Subcommittee on Human Resources, Committee on Government
Reform and Oversight, House of Representatives

For Release on Delivery
Expected at 10:00 a.m.
Thursday, June 5, 1997

BLOOD SAFETY - ENHANCING
SAFEGUARDS WOULD STRENGTHEN THE
NATION'S BLOOD SUPPLY

Statement of Bernice Steinhardt, Director
Health Services Quality and Public Health Issues
Health, Education, and Human Services Division

GAO/T-HEHS-97-143

GAO/HEHS-97-143T


(108335)


Abbreviations
=============================================================== ABBREV

  AIDS - acquired immune deficiency syndrome
  CBER - Center for Biologics Evaluation and Research
  CFR - Code of Federal Regulations
  FDA - Food and Drug Administration
  HIV - human immunodeficiency virus
  HHS - Department of Human and Health Services
  HTLV - human T-lymphotropic virus
  HTLV-II - human T-lymphotropic virus type II

BLOOD SAFETY:  ENHANCING
SAFEGUARDS WOULD STRENGTHEN THE
NATION'S BLOOD SUPPLY
============================================================ Chapter 0

Mr.  Chairman and Members of the Subcommittee: 

It is a pleasure to be here this morning to discuss our examination
of the safety of the nation's blood supply.  Donors give
approximately 14 million units of whole blood and 12 million units of
plasma annually.  As many as 4 million patients receive transfusions
of whole blood components and millions more receive plasma products
each year.  Since the human immunodeficiency virus (HIV) was
introduced into the U.S.  blood supply in the early 1980s, the
benefits of a potentially life-saving transfusion have had to be
weighed against the risks of acquiring this deadly disease through
blood transfusion. 

Widespread concern about the safety of the blood supply has led to
many positive changes in the way blood is collected, processed, and
transfused.  In testimony on July 28, 1993, before the Subcommittee
on Oversight and Investigations of the House Committee on Energy and
Commerce, the Commissioner of the Food and Drug Administration (FDA)
outlined five overlapping "layers of safety" that provided a
framework for regulating and monitoring the blood supply industry: 
(1) donor screening, (2) donor deferral registries to list unsuitable
donors, (3) viral testing, (4) quarantining blood until tests and
control procedures have established its safety, and (5) monitoring
facilities and investigating adverse incidents to ensure that
deficiencies are corrected. 

While the blood supply is very safe, no amount of federal regulation
can entirely eliminate blood transfusion risks because of the
biological nature of the product itself.  Increasingly sophisticated
tests are shortening the time between infection and detectability of
infection in the blood.  Blood donated during this interval, known as
the window period, is the leading cause of infected blood remaining
in the blood supply.\1

Improved viral tests will continue to close this gap, but the window
period is not likely to disappear completely. 

My statement today is based on our two reports on the blood supply
issued in February 1997.\2 In these reports, we assessed the current
risks of transfusion, evaluating the content and quality of data
collected to assess these risks.  We also evaluated the FDA's layers
of safety and their ability to ensure the safety of the nation's
blood supply. 

In summary, our analysis of current risks from transfusion showed
that, while the nation's blood supply is safer today than at any time
in recent history, some risk remains, even if all the safeguards
available work perfectly.  We also found several vulnerabilities and
gaps in current procedures which, if eliminated, would provide
greater assurance of safety for the nation's blood supply.  The most
serious of these problems follow: 

  Not all donors who test positive for certain viruses are notified,
     which means that they can attempt to donate again and also may
     go without treatment. 

  Similarly, not all recipients of virally contaminated blood are
     notified, which may keep them from seeking treatment and also
     allow them to transmit the disease. 

  Blood facilities are not required to remove from their inventory
     blood from donors who have subsequently tested positive for
     viral infections. 

  Unlicensed blood facilities that, together, produce 10 percent of
     the nation's blood do not have to submit to FDA error and
     accident reports that may signal the need to recall potentially
     contaminated units of blood. 

  FDA's investigations of error and accident reports that warrant a
     recall take a long time and increase the risk that units will
     have been transfused before a recall is accomplished. 

  Finally, FDA's inspections of blood facilities are inconsistent in
     focus, scope, and documentation. 

Our reports contained a number of recommendations to the Secretary of
Health and Human Services to eliminate these weaknesses in the
quality assurance system for the blood supply. 


--------------------
\1 The window period is the time from infection to the point at which
currently licensed test kits can ascertain antibodies or antigens to
certain viruses tested for by blood facilities. 

\2 Blood Supply:  Transfusion-Associated Risks (GAO/PEMD-97-2, Feb. 
25, 1997) and Blood Supply:  FDA Oversight and Remaining Issues of
Safety (GAO/PEMD-97-1, Feb.  25, 1997). 


   TRANSFUSION-ASSOCIATED RISKS
---------------------------------------------------------- Chapter 0:1

At this time, I would like to tell you more about our analysis of the
current risks of transfusion-associated complications from blood,
assuming all layers of safety are working properly--that is, blood
from donors who were properly screened, whose names were checked
against a deferral registry, whose viral test results were negative,
and so on. 

In conducting our analysis, we reviewed current data and the
scientific literature as well as interviewed government and industry
epidemiologists.  We then compared our final estimates on risks from
blood transfusions with data on risks from other health-related
causes.  We included risks from eight viruses, various bacteria, one
parasite, and four complications of transfusion itself.\3 When we
encountered differing estimates of risks from research that we
considered equally valid, we chose the higher estimate. 

We found that the blood supply is safer today than at any time in
recent history.  Nevertheless, blood is a biological product, and
some risk remains.  Eight of every 10,000 donated units carry some
kind of potentially serious risk to the recipient, including allergic
reactions, bacteria, reactions to incompatible blood, and viruses. 
We calculated that 4 of every 1,000 patients who receive the average
transfusion of 5 units of blood are at risk of receiving a
contaminated unit and thus may be exposed to conditions with the
potential for the development of serious (that is, chronic,
disabling, or fatal) outcomes at some point in the future.  We
believe this risk is small considering that as many as 50 percent, or
500, of the 1,000 recipients would be at serious risk of dying
immediately if they did not receive transfusions. 

Moreover, not all recipients of a contaminated unit acquire the
disease it contains.  And, many recipients die soon after transfusion
from the underlying condition for which the blood was prescribed. 
Finally, the likelihood that a patient will develop chronic disease
or die is small for some diseases that are transmitted by
transfusion.  We determined that the overall risk of developing
chronic disease or dying as a direct result of a blood transfusion is
about 4 in 10,000, which translates into about 1,525 of the 4 million
patients who receive transfusions each year.  Thus, if all the
safeguards are working properly, the risks are relatively small and
are certainly far outweighed by the benefits. 

Because risks should never be discussed out of context, we sought to
determine whether these transfusion risks were small or large by
comparing them with other known health-related risks.  The risks to
blood transfusion recipients are considerably smaller than the risk
of dying as a direct result of surgery, the risk that a hospital stay
will result in death or chronic disability, the risk of suffering a
serious injury from hospital drug therapy, and the risk of developing
an infection of unknown cause in intensive care. 


--------------------
\3 The viruses included were hepatitis A, hepatitis B, hepatitis C,
HIV-1 and HIV-2, human T-lymphotropic virus type I (HTLV-I) and type
II (HTLV-II), and non-ABC hepatitis.  The parasite-transmitted
disease included was Chagas', and the transfusion complications were
ABO incompatibility, acute lung injury, allergic reaction, and
circulatory overload. 


      RISKS FROM PLASMA PRODUCTS
-------------------------------------------------------- Chapter 0:1.1

The risk estimates I have just presented are for whole blood products
from unpaid donors, which account for about half of all donations. 
The remaining blood products are plasma products, such as immune
globulins or clotting factors, which are usually obtained by
commercial facilities from paid donors.  Because only limited data
are available concerning the risks posed by plasma products, we were
unable to include plasma derivatives in our analysis of risks. 
However, because the ways in which plasma products are manufactured
differ from the way whole blood products are prepared, and because
these products are used differently, it may be worth highlighting
some of these features to try to understand the nature, if not the
full extent, of risks associated with this sector of the blood
supply. 

More than 40 million hospital patients use plasma products each year. 
Plasma is the liquid portion of blood, containing nutrients,
electrolytes (dissolved salts), gases, albumin, clotting factors,
hormones, and wastes.  Many different components of plasma are used
for purposes that range from treating the trauma of burns and surgery
to replacing blood elements that are lacking as a result of a disease
such as hemophilia. 

In the 1980s, before the etiology of HIV transmission was understood,
many hemophilia patients used plasma products infected with HIV, and
63 percent of all hemophilia patients in the United States became
infected as a result.  Many more contracted hepatitis B and hepatitis
C.  Since the introduction of antibody tests and heat treatments and
solvent-detergent washing processes for inactivating and removing
viruses, however, the transmission of disease has been considerably
reduced. 

Current techniques appear effective for protecting against the
transmission of HIV, hepatitis C, and hepatitis B.  But certain
viruses that are not surrounded by a fatty envelope--such as
hepatitis A and parvovirus--are not inactivated by current
techniques.  Moreover, different manufacturers producing similar
products may or may not use these techniques.\4 The extent to which
current manufacturing techniques will be effective against unknown
pathogens that could enter the blood supply is not known. 

Despite the evidence that viral inactivation and removal processes
improve the safety of plasma products, the fact remains that the paid
plasma donor pool has higher rates of viral infectivity than the
volunteer whole blood donor pool.  Unlike whole blood, plasma is
typically collected from paid donors in a commercial setting.  In
1978, FDA required that each blood unit be labeled as either
volunteer or paid.  In the regulations, FDA concluded that paid blood
donors were more likely to transmit hepatitis to recipients than were
volunteer donors.  FDA's conclusions were based on research evidence
showing higher rates of hepatitis in commercial donors and in
recipients of paid donor blood as well as evidence showing that the
elimination of commercial blood resulted in substantially fewer cases
of posttransfusion hepatitis.  While the commercial donor pool for
whole blood is all but nonexistent in the United States today, the
plasma industry continues to rely on paid donors to supply the raw
plasma for further manufacturing into plasma derivatives. 

We were unable to obtain national data on the viral test positivity
rates among paid plasma donors compared with those of volunteer blood
donors.  We did, however, find several sources of information
pertaining to this issue.  First, we found that California requires
the reporting of initial and confirmed HIV prevalence rates for both
blood banks and plasma collection centers.  Figure 1 shows that the
confirmed HIV prevalence rates per 100,000 commercial plasma
donations in California have decreased in recent years but remain
substantially higher than those same rates for volunteer whole blood
donations. 

   Figure 1:  Quarterly Confirmed
   HIV Prevalence Rates for
   Donations in California, Fiscal
   Years 1989-94

   (See figure in printed
   edition.)

Note:  These rates are reported Western Blot-confirmed HIV prevalence
rates per 100,000 commercial plasma donations and volunteer whole
blood donations. 

Source:  California Department of Health Services, Office of AIDS,
HIV/AIDS Epidemiology Branch, Sacramento, California, August 1995. 

Unlike whole blood donors, who cannot donate blood more often than
once every 8 weeks, plasma donors can donate twice a week.  As a
result, fewer plasma donors are needed to collect 100,000 units. 
Moreover, several plasma units could be donated during a window
period, whereas it is unlikely that more than one whole blood unit
could be donated in a window period. 

We also analyzed the clinical data that plasma manufacturers
submitted to FDA during the approval process for several viral tests. 
The test-positive rates for commercial plasma donors were
substantially higher than those of volunteer whole blood donors,
ranging from about 2 to 20 times higher on the different tests. 

While most commercial plasma donors are healthy and free of disease,
monetary incentives such as those offered by commercial
plasma-collection centers may be tantalizing to some of those who are
known to be at risk for infectious diseases, such as intravenous drug
users and prostitutes.  Screening questions address these risk
behaviors, but there is no definitive way to screen out all risky
donors, and current tests may not be sufficient to catch all infected
units. 

Newly emerging and yet unknown viruses often enter the population
through high-risk individuals.  Viral antibody tests may not yet
exist for these new viruses, and current viral inactivation and
removal techniques may be ineffective for them.  Moreover, one
infectious donation can contaminate an entire pool of as many as
60,000 units.  Without national data on the differences in prevalence
and incidence rates between paid and volunteer donors, it is not
possible to draw firm conclusions about potential risks posed by
plasma derivatives.  Such data would be valuable because they could
be used to monitor the blood industry in its entirety. 


--------------------
\4 In December 1994, FDA notified manufacturers of immune globulin
products that it would begin testing for hepatitis C in all products
that had not undergone a validated virus inactivation or removal
process.  The products affected by this policy include Rho(D) immune
globulin for Rh-negative pregnant women and specific immune globulins
for hepatitis B; tetanus; and varicella-zoster, the agent that causes
chicken pox.  No new cases of hepatitis C transmission by these
intravenous products have been reported to date.  A similar product,
immune globulin for intramuscular administration, is not virally
inactivated.  Although no cases of hepatitis C transmission by
intramuscular administration of immune globulin have ever been
reported, concerns have been raised about this product, and FDA
allows only the manufacturing lots that have been tested for
hepatitis C to be distributed.  HIV is a delicate virus that is
readily inactivated.  No cases of HIV transmission by plasma products
inactivated according to current standards have been reported. 


   FDA OVERSIGHT AND REMAINING
   ISSUES OF SAFETY
---------------------------------------------------------- Chapter 0:2

To test whether blood supply safeguards are working, we examined
FDA's layers of safety and found vulnerabilities throughout,
including problems in the areas of donor screening, notification,
postdonation information, recalls, and FDA standards and
inspections.\5 These vulnerabilities are summarized in the appendix. 


--------------------
\5 We limited the scope of our investigation to policies and
procedures that were current in 1994.  Thus, we did not examine
problems and consequent policy changes of the mid-1980s as a result
of the discovery that HIV can be transmitted via blood transfusion. 
Nor did we examine the patterns of violations by individual
facilities.  The focus of our work was the general policies and
procedures in place to help ensure the safety of the blood supply. 


      DONOR SCREENING
-------------------------------------------------------- Chapter 0:2.1

Donor screening, the first layer of safety, is designed to prevent
the donation of blood by people who have known risk factors for
disease transmission or are not in good health.  High-risk donors,
those whose blood may pose a health hazard, are encouraged to exclude
themselves.  All potential blood donors must answer a series of
behavioral and medical questions.  If any one answer indicates high
risk, the prospective donor is not allowed to donate.  If the
questions are answered truthfully, they isolate about 90 percent of
people whose risk of having HIV is too recent for their bodies to
have produced sufficient antigen or antibodies that would be detected
by viral screening tests.\6

We found two potential vulnerabilities in the area of donor
screening.  First, while questioning and screening donors about their
behaviors and medical history is important in maintaining a safe
blood supply, studies have shown that the style and content of
history taking may influence the accuracy and completeness of donor's
answers.  The American Association of Blood Banks has a comprehensive
and readily available uniform donor history questionnaire that, if
adopted by more facilities, could strengthen donor screening
procedures.  Second, the amount of privacy for screening donors
varies across blood facilities.  A lack of privacy during donor
screening inhibits forthright communication. 

The importance of screening donors with validated questionnaires in a
private environment is underscored by a study published after our
reports were issued of 35,000 blood donors who completed a mail
survey 4 to 8 weeks after their most recent blood donation.\7 A total
of 186 per 10,000 donors (1.9 percent) reported a deferrable risk
that was present at the time of their donation, and 39 per 10,000
donors (0.4 percent) reported having engaged in behaviors that should
have resulted in deferral within the 3 months prior to donation. 
Further refinement of the donor qualification process could help
deter these potentially risky donors from donating blood. 


--------------------
\6 Antibody tests detect antibodies that the human body produces in
its immune response to a virus, whereas antigen tests detect a
component of the actual virus.  Because it takes time to develop
antibodies, antigen tests detect infection earlier than antibody
tests. 

\7 Alan E.  Williams and others, "Estimates of Infectious Disease
Risk Factors in U.S.  Blood Donors," Journal of the American Medical
Association, 277:12 (1997), pp.  967-72. 


      NOTIFICATION
-------------------------------------------------------- Chapter 0:2.2

At both the deferral and testing layers, blood facilities have an
opportunity, and sometimes a requirement, to notify donors as well as
recipients of indications of disease.  We found two areas of concern
related to notification.  Not all blood facilities notify donors that
they have tested positive on a viral screening test and that they are
deferred from donating again.\8 FDA recommends notification of donors
deferred for HIV only.  While the blood is not used in cases in which
test results are positive, this does not ensure that these donors
will not attempt to donate at another site; neither does it prompt
them to change behaviors or seek treatment so that they do not
transmit the disease to family members or others. 

Also contributing to this problem is the fact that facilities vary in
the extent to which they perform confirmatory or supplementary tests
on blood that has repeatedly tested reactive on initial screening
assays.  FDA only requires confirmatory testing of HIV-positive
units.  Units repeatedly reactive for other viruses do not always
have confirmatory tests performed on them, and confirmatory tests for
some viruses have not been developed or licensed by FDA.  Thus,
facilities that do not perform such tests cannot adequately inform
donors about their disease status, even if they notify donors that
they are deferred. 

Facilities also vary in their policies for notifying recipients who
have received blood from donors who later test positive for viruses
and for conducting lookback, that is, tracing and removing units from
implicated donors that remain in inventory.  FDA requires these
practices for HIV and recommends--but does not require--quarantine
and destruction of units in inventory from donors who subsequently
have repeatedly reactive tests for hepatitis B, hepatitis C, and
HTLV.  FDA has made no recommendations about notifying recipients who
may have received blood infected with these other viruses. 

Not notifying these recipients poses a potential public health
problem.  Using hepatitis C as an example, we found that, although
the mechanisms of secondary transmission are not well established,
some secondary transmission of hepatitis C does occur.  The Centers
for Disease Control and Prevention has issued guidance for infected
people that includes recommending protected sex for individuals with
multiple partners and the avoidance of sharing common household
articles, such as razors and toothbrushes.  Furthermore, abstinence
from alcohol is strongly recommended for infected people because
alcohol intake results in more liver disease and increases the risk
of liver cancer.  Although medical therapies are not yet 100-percent
effective, clinical trials for alpha interferon therapy show that 23
percent of patients achieved a long-term remission at the end of
treatment.  We believe recipients of hepatitis C-infected blood
should have the right to decide with their physicians whether medical
therapy is indicated for their disease.  Moreover, should a more
effective therapy arrive in the future, recipients who are not
notified today would likely be lost to follow-up. 


--------------------
\8 Screening tests are conducted for hepatitis B by testing for
surface antigen (an indication of active virus) and antibody to core
(an indication of resolving or past infection and a surrogate marker
for high-risk behavior, such as intravenous drug use); for hepatitis
C by antibody test; for HIV by antibody and antigen tests; for HTLV-I
by antibody test; and for syphilis by serological test.  Increasingly
sophisticated tests are closing the time between infection and
detectability of infection in the blood. 


      POSTDONATION INFORMATION
-------------------------------------------------------- Chapter 0:2.3

Another critical layer of safety is the quarantining of blood for a
period of time following donation during which additional information
and test results may lead to the decision that the blood is
unsuitable for use.  For example, donors may provide information
after donating that would have excluded them from donating had it
been known at the time of donation.  Sometimes donors call to report
relatively minor issues such as having developed a cold; other times,
donors call to say that they engage in behaviors (such as intravenous
drug use) that put them at serious risk of disease; still other
times, donors report at a subsequent donation attempt that they
engage in behaviors that put them at serious risk of disease.  If
such postdonation information is received after a unit is made
available for distribution, the blood facility must submit this
information as an error and accident report--a type of report that a
facility must file with FDA whenever it discovers a mistake that
affects the safety, purity, or potency of blood products. 
Postdonation information accounted for about 3,800, or more than
one-third, of all error and accident reports in fiscal year 1994. 

The preponderance of errors and accidents related to postdonation
information is a concern.  It could indicate that the system is
working properly or that FDA should more clearly define what is to be
reported.  The large proportion of errors and accidents discovered as
a result of postdonation information also calls into question the
adequacy of screening processes.  For example, 65 percent of the
error and accident reports related to postdonation information
stemmed from information obtained at a subsequent donation. 

While we cannot explain the differences, we found far fewer
postdonation error reports from plasma centers than from licensed
whole blood facilities:  Whole blood facilities' reporting rate was
135 times higher, although both collect approximately the same number
of units each year.  Since data show higher prevalence rates of HIV
and perhaps other diseases at plasma centers, as we pointed out
earlier, this appears to be an area where more information is
warranted. 


      RECALLS
-------------------------------------------------------- Chapter 0:2.4

As the final layer of safety, blood facilities are obligated to
monitor and investigate errors and accidents in their procedures, to
audit their systems, and to correct deficiencies.  As explained
earlier, if an error or accident results in a potentially
contaminated unit of blood being made available for distribution,
licensed facilities (both whole blood and plasma) are required to
report the incident to FDA.  Unlicensed facilities are requested to
voluntarily report such incidents. 

Once a facility reports an error or accident to FDA's Center for
Biologics Evaluation and Research (CBER), depending on the severity
of the incident, FDA's district field office located nearest the
facility evaluates it and may recommend a recall.  Most recalls are
initiated by the responsible establishment and often are completed
before FDA learns of them.  Recalls are voluntary; while FDA may
prompt a firm to initiate a recall, this occurs in only 25 percent of
recalls.  In egregious cases, such as those posing an imminent threat
to the public where the blood establishment resists initiating a
recall, FDA has the authority to initiate product recalls but has
never done so for blood products.  CBER's role is to determine that
an unsuitable product should be recalled if the establishment has not
already done so and to classify the recall based on a health hazard
evaluation to establish the level of FDA follow-up required to ensure
that the public is protected. 

Only licensed facilities are required to submit error and accident
reports to FDA.  Although unlicensed facilities are asked to
voluntarily submit their reports, FDA's annual summaries suggest that
unlicensed facilities may be underreporting.  Our analysis of FDA's
summary for fiscal year 1994 found that unlicensed facilities submit
only 12 reports for every 100,000 units of blood they collect,
compared with 82 reports per 100,000 units for whole blood facilities
(see fig.  2).  This means that unlicensed facilities submit only
about 1 percent of the reports, although they account for 10 percent
of the blood supply.  While plasma centers are required to submit
error and accident reports, they also report at rates much lower than
licensed whole blood facilities, despite collecting equivalent
amounts of blood products.  Moreover, 39 percent of the error and
accident reports that CBER received from plasma centers were sent
forward to the districts to be reviewed for potential product
recalls, as compared with only 5 percent of reports submitted by
licensed whole blood facilities. 

   Figure 2:  Total Error and
   Accident Reports by Facility
   Type, Fiscal Year 1994

   (See figure in printed
   edition.)

Source:  GAO's analysis of FDA's Annual Summary for fiscal year 1994. 

Unlicensed facilities also submit fewer error and accident reports in
situations that end in product recalls.  In roughly two-thirds of the
recalls in 1994, a report was submitted before the district office's
recommendation for recall:  Nearly all of these reports came from
licensed facilities, including plasma centers.\9 More than 70 percent
of licensed facilities submitted a report before recall, but only 17
percent of unlicensed facilities did this.  Given that these reports
are one way of alerting FDA to the need for an immediate recall, we
believe that underreporting by unlicensed facilities is a serious
problem. 

In those cases in which facilities are reporting, the Department of
Health and Human Services' (HHS) Inspector General's Office has found
that timeliness is a problem.\10 For a random sample of 163 reports
from October 1992 to April 1993, the time between the date when a
blood facility detected an error or accident and the date when this
information was submitted to FDA ranged from less than 1 month to
more than 1 year, the average being a little over 4 months.  While 14
percent of reports were submitted within 1 month, 13 percent were
reported 6 months or more after the error was detected. 

Further, we found that timeliness of FDA actions in response to
reports is also a problem.  Our analysis of FDA's recall database
showed that in 60 percent of cases, 7 months or more elapsed between
the time of report submission and the district office's
recommendation to CBER that a recall should be considered.  The
average time for CBER review was 9 weeks, but reviews sometimes took
as long as a year.  The total time from report submission to recall
confirmation and public announcement ranged from a little over 1
month to 2-1/2 years, with an average of nearly 9-1/2 months; in 70
percent of cases, the time was 7 months or more (see fig.  3). 

   Figure 3:  Time Elapsed From
   Error and Accident Detection to
   Recall Confirmation, October
   1992-April 1993

   (See figure in printed
   edition.)

Note:  Numbers may not sum to 100 percent because of rounding. 

Source:  GAO's analysis of FDA Recall Action Database. 

We found no significant differences in FDA's processing time based on
the severity of the case.  That is, more serious cases were not
processed faster than less serious ones.  Given the long time FDA
takes to go through its formal recall process, blood product safety
could be compromised.  Clearly, the longer it takes to initiate a
recall, the more likely it is that all the product will have already
been transfused. 


--------------------
\9 Our statistical analysis determine that this difference between
licensed and unlicensed facilities was highly significant (t = -8.96,
p <.0001). 

\10 Office of Inspector General, HHS, Reporting Process for Blood
Establishments to Notify the Food and Drug Administration of Errors
and Accidents Affecting Blood, A-03-93-00352 (Washington, D.C.:  HHS,
May 1995). 


      FDA STANDARDS AND GUIDELINES
-------------------------------------------------------- Chapter 0:2.5

FDA communicates its requirements through the Code of Federal
Regulations and its policies and recommendations through memorandums
and letters, compliance manuals and the compliance program,
compliance policy guides, and a guide for blood facility inspections. 
The requirements in the Public Health Service Act; the Food, Drug,
and Cosmetic Act; and the C.F.R.  are the only mandatory
requirements. 

We found substantial confusion in the industry on the distinction
between FDA regulations and guidance, potentially leading to
different interpretations and applications of FDA's requirements and
recommendations.  As part of our review, we conducted a survey of 45
full-service blood facilities.\11 Many of our survey respondents told
us they were unclear about which statements had to be followed and
which were only FDA recommendations.  Respondents also noted that FDA
inspectors sometimes filed reports on significant infractions--forms
483--on the basis of FDA recommendations, that the regulations should
be updated to incorporate current memorandums, and that the language
in the memorandums should be clarified to indicate which actions are
required and which are recommended.\12

A 1995 Institute of Medicine study on blood safety issues recommended
that "when issuing instructions to regulated entities, FDA should
specify clearly whether it is demanding specific compliance with
legal requirements or is merely providing advice for careful
consideration."

The issue has practical implications.  The law explicitly requires
FDA to prescribe standards for insuring purity, potency, and safety
of blood products.  However, regarding HTLV testing, FDA has issued
memorandums on such procedures; its regulations do not refer to HTLV
testing at all.  Thus, a facility could be licensed and yet view
testing for HTLV as only a recommendation and not a requirement. 
Nevertheless, not testing for HTLV could directly affect the safety
of blood products. 

To its credit, FDA has historically issued memorandums to give the
industry immediate feedback on its positions on new issues.  However,
guidelines and memorandums issued for expedience appear to rarely
move into the formal regulatory process.  While blood facilities
often incorporate recommendations into standard operating procedures,
the lack of a public comment period--as is required in the formal
rulemaking process involved in setting regulations--gives blood
facilities no opportunity to address important implementation issues
and could lead to inconsistent policies in the industry. 


--------------------
\11 By "full-service," we mean those facilities that conduct the full
range of blood collection, processing, and distribution (including
viral testing).  The response rate to our survey was 100 percent. 

\12 An FDA inspector who identifies significant infractions that
could affect blood safety files a form 483, noting the objectionable
conditions. 


      FDA INSPECTIONS
-------------------------------------------------------- Chapter 0:2.6

We found several problems in FDA's inspection process in four broad
categories:  the use of inspection reports, the timing of
inspections, the completeness of inspection reports, and the
consistency of inspection reporting.  FDA inspects blood facilities
every 2 years.  Facilities that have received a warning letter or
have been found deficient in inspections within the past 2 years may
be inspected annually until they pass two consecutive inspections
without significant observations.  Inspectors file an establishment
inspection report with FDA at the close of the inspection, which
descriptively narrates the activities covered in the inspection and
any problems identified.  Observations of potentially unsafe
conditions are filed on a form 483 and discussed with management of
the facility. 

We were told by FDA that it reviews all inspection reports.  However,
we found that FDA conducts no systematic statistical analyses of
inspection reports or forms 483.  Without collating, synthesizing,
analyzing, and evaluating these data, FDA has no means of assessing
overall national compliance, assessing trends by type of facility,
identifying the problems of different types of blood facilities, or
evaluating the effects of policy changes on implementation rates.  By
performing these types of statistical analyses, FDA could obtain
information on different rates of form 483 observations among
district offices, rates of observations by type of activity (for
example, donor screening, donor deferral, and viral testing), and
rates among types of facilities.  We conducted such an analysis,
discussed below, which illustrates the feasibility and importance of
this task. 

We obtained inspection reports and form 483 reports of inspection
observations on a nationally representative sample of blood
facilities.  FDA's own requirement is to inspect blood facilities
every 2 years, or more often if significant violations have been
detected.  However, of the 373 blood facilities in our sample, 45 (12
percent) had not been inspected in more than 2 years.  Because our
sample represented all blood facilities in the nation, we could
project that 348 of the 2,900 registered blood facilities (12
percent) may not have been inspected within the past 2 years. 

We also found problems with the completeness of inspection reports. 
We examined each facility in our sample for whether the inspection
report indicated that a particular function (such as viral testing)
had been examined.  For the purpose of our analysis, if it was
mentioned at all in the report, we considered it to have been
examined.  If it was not mentioned anywhere in any way, we considered
that one could not determine whether the area had been examined. 

For the time period when checklists were required, we found that 40
of 224 inspections (18 percent) that should have included an
inspection checklist did not have one.\13 In many instances, we were
unable to determine whether procedures relating to donor screening,
deferral, collection, routine testing, viral testing, postdonation
information, labeling, quarantining, storage, and "machines" were
examined at all in the individual inspections.  In fact, for all the
areas in our analysis that FDA should have inspected, we could not
find indications that it did so in 33 percent (963 of 2,957 areas). 
Further, we were able to determine in only half of all reviewed
reports that inspections covered all activities necessary to ensure
compliance. 

FDA's current policy is for the inspectors to list on the inspection
report only areas that were not covered.  That is, when an inspector
notes on the report that the inspection was undertaken within a
specific compliance program, this means that all blood banking
practices covered in the compliance program have been examined.  We
found that this policy is unreliable in ensuring that activities not
covered during the inspection are, in fact, noted on the report. 
Moreover, without detailed information, FDA supervisors or subsequent
inspectors cannot determine what blood banking processes have been
examined in an inspection. 

For example, at a blood facility inspected in 1994, an inspector
found that no lookback procedures had been followed in several cases
of reported HIV-positive donors identified since 1990.  When we
examined the inspection report for this facility for the inspection
that took place in 1993, we found no indication that lookback
procedures were not being followed.  This means either that the 1993
inspection examined lookback procedures and did not find that they
had not been carried out since 1992 (according to the 1994
inspection) or that lookback procedures were not observed in the 1993
inspection and this was not noted on the inspection report, which is
FDA's stated policy. 

As a further measure of the comprehensiveness of inspections, we
asked the 45 full-service blood facilities in our survey to what
extent FDA examined standard operating procedures in 12 separate
areas in their last inspection.  In every area except deferral, more
than half the respondents indicated that FDA examined standard
operating procedures only to a moderate extent or less.  Similarly,
the respondents reported that FDA does not observe or otherwise
examine firsthand major activities in many areas.  More than 20
percent reported little or no FDA observation of six different areas. 
Furthermore, 35 percent of the respondents indicated that FDA
evaluated the existence and suitability of only half or fewer of the
critical control points their facilities had in place to ensure
safety, purity, and potency. 

Finally, we have concerns relating to the consistency of inspection
reporting.  We found significant disparities in inspection reporting
across the eight FDA districts we examined.  For example, more than
21 percent of form 483 observations related to labeling in one
district but only 2 percent in another.  We also found statistically
significant differences between districts in the issuance of forms
483.  In particular, one district issued forms 483 to only 20 percent
of inspected facilities, compared with a range of 42 to 52 percent
among the districts most likely to issue a form 483.  Districts
differed in the types of activities that warranted forms 483.  Why
observations are issued inconsistently is not clear.  Either
different districts have different problems, or different districts
interpret FDA policy differently.  Neither we nor FDA can say which
is the case.  Yet 27 percent of our survey respondents reported that
they do not know what to expect from one inspection to the next; what
is acceptable to one inspector, they say, may be an unsafe condition
to another.  And while respondents reported that their most recent
inspection team was knowledgeable about blood banking terminology and
technology, 45 percent reported a wide variation among inspectors. 


--------------------
\13 In September 1994, FDA replaced the checklist with a
systems-based guide. 


   CONCLUSIONS AND RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:3

While FDA, together with industry, has made great strides in
improving the nation's blood supply since the recognition of the
risks posed by HIV, we believe that eliminating the vulnerabilities
we identified would enhance the safety of blood products. 

Therefore we have recommended that the Secretary of Health and Human
Services take the following actions: 

  Require that blood facilities notify all donors who are permanently
     deferred (not just those who test positive for HIV) that they
     have been deferred and the medical reasons for their deferral,
     so that they do not attempt further donation and can seek
     further medical care if they desire. 

  Require confirmatory testing of all repeatedly reactive viral test
     results for which there is a licensed confirmatory test, in
     order for blood facilities to be able to properly counsel donors
     as to their disease status. 

  Require that patients be notified when they have been transfused
     with blood from a donor whose subsequent donations were found to
     be positive by confirmatory testing for any virus for which a
     confirmatory test is available, not just for HIV.  We note that
     the reasonable time period for tracing back units to recipients
     varies with each virus, and decisions should be made in
     consultation with the blood industry. 

  Require lookback to identify and remove units from implicated
     donors that remain in inventory in situations in which those
     donors' subsequent donations are found to be positive by
     confirmatory testing for any virus for which a confirmatory test
     is available, not just for HIV. 

  Require unlicensed facilities to report all errors and accidents. 

We have recommended that the Secretary take the following additional
actions: 

  Publish in the form of regulations the guidelines that FDA deems
     essential to ensure the safety of the blood supply and require
     that FDA clarify its position on the extent to which facilities
     must adopt guidelines and memorandums in order to remain in
     compliance. 

  Correct problems that we have identified in FDA inspection
     processes.  FDA should perform statistical analyses of
     inspection reports, ensure that all blood facilities are
     inspected in a timely fashion, develop policies for the
     inspectors to list on inspection reports the activities they
     observe, and publish better guidance to inspectors on the types
     of activities that warrant reports on deviations and warning
     letters. 

FDA has been aware of a number of these problems for several years
and has initiated some actions to address them.  In other cases, the
agency has said that our recommendations would be too costly or
unnecessary. 

We remain convinced, however, that if all the improvements we
identified are made, the American public will be better assured that
the blood supply is as safe as possible given the current state of
technology and medical knowledge.  Continued safety depends on the
scientific and medical communities' vigilance in detecting and
identifying any new threats to the supply. 

This concludes my prepared statement, Mr.  Chairman.  I will be happy
to respond to any questions that you or Members of the Subcommittee
may have. 


REMAINING VULNERABILITIES IN THE
LAYERS OF BLOOD SAFETY
==================================================== Appendix Appendix

------------------------------  ---------------------------------------------------------
Donor screening                 --The style and content of history-taking questionnaires
                                may influence the accuracy and completeness of donors'
                                answers.
                                --Lack of privacy at some facilities may inhibit
                                forthright communication.

Notification                    --Lack of universal donor deferral notification could
                                create public health problems.
                                --Lack of universal confirmatory testing of donors
                                testing repeatedly reactive on initial screening assays
                                precludes facilities from having complete information on
                                disease status to use in notifying donors and
                                recipients.
                                --Except for HIV, recipients who have received
                                potentially infectious blood do not have to be notified,
                                and blood facilities do not have to trace and remove
                                units that remain in inventory.

Postdonation information        --Many errors and accidents are discovered as a result of
                                postdonation information that would have excluded the
                                donor had it been known at donation.
                                --Plasma centers report proportionately fewer
                                postdonation errors and accidents than licensed whole
                                blood facilities, despite being subject to the same
                                reporting requirement and collecting equivalent amounts
                                of blood.

Recalls                         --Only licensed facilities are required to report.
                                --Plasma centers report proportionately fewer errors and
                                accidents in all areas, despite being subject to the same
                                reporting requirement and collecting equivalent amounts
                                of blood.
                                --Report submissions and subsequent FDA investigations
                                are not always timely.

FDA standards and inspections   --FDA guidance to blood facilities is often ambiguous.
                                --FDA does not perform statistical analyses on inspection
                                reports and forms 483 and therefore cannot assess
                                compliance trends.
                                --Some facilities are not inspected within FDA-
                                established timeframes.
                                --Inspection reports are often incomplete.
                                --Differences exist in form 483 observations among FDA
                                districts, including disparities in what actions
                                constitute need for further action.
-----------------------------------------------------------------------------------------

*** End of document. ***