       Document 0673
 DOCN  M94A0673
 TI    Assessing health service interventions in HIV-disease.
 DT    9412
 AU    Beck EJ; Academic Department of Public Health, St. Mary's Hospital
       Medical; School, Imperial College, UK.
 SO    Annu Conf Australas Soc HIV Med. 1993 Oct 28-30;5:46 (abstract no. SE1).
       Unique Identifier : AIDSLINE ASHM5/94348982
 AB    AIM: To assess the effectiveness of health service interventions in
       HIV-disease, pathophysiological indices need to be interpreted within
       the context of service utilisation and specific treatment received. This
       interrelationship will be demonstrated in terms of outcome of first AIDS
       defining (index) episode of Pneumocystis carinii pneumonia (PCP),
       longterm survival from time of index episode of PCP and changing
       patterns of service utilisation. METHOD: Pathophysiological indices and
       health service utilisation data were collected retrospectively from
       casenotes of patients with index episode of PCP treated at St. Mary's
       Hospital, 1983-90. Univariate analysis identified biomedical and service
       utilisation indices associated with episode outcome (death/survival) and
       logistic regression was used to integrate these variables. Life
       regression and life test analyses were used to relate index episode to
       two-year survival. RESULTS: 159 (74%) of first time episodes during the
       study period were index episodes. Case-fatality decreased significantly
       over time as did A-a gradient on admission. Conversely, the time
       interval between diagnosis of HIV infection and PCP (HIV-PCP) increased:
       duration HIV-PCP was strongly associated with number of outpatient
       visits and lung function tests prior to index episode, while a
       significant inverse relationship existed between A-a gradient and
       duration HIV-PCP, outpatient visits and lung function test respectively.
       Logistic regression demonstrated that A-a gradient was inversely related
       with episode outcome while haemoglobin level, duration HIV-PCP and
       lymphocyte count were directly associated with episode outcome. Type of
       treatment received during the admission was also related to severity of
       PCP on admission. Two-year survival improved significantly over the
       study period. Life regression demonstrated lymphocyte count and A-a
       gradient of index episode to be significantly related to longterm
       survival. When controlling for lymphocyte count and A-a gradient,
       haemoglobin, duration HIV-PCP and year of diagnosis were also
       significantly associated with longterm survival. DISCUSSION: Earlier
       presentation, diagnosis and treatment of patients with and index episode
       of PCP reduced case-fatality and improved longterm survival. These
       improvements occurred independently of zidovudine therapy or primary PCP
       prophylaxis. Only when pathophysiological parameters are interpreted
       within the context of service utilisation can the effectiveness of
       health service interventions be assessed. Effective interventions then
       need to be assessed in terms of their efficiency and acceptability in
       order to determine HIV-related service provision.
 DE    AIDS-Related Opportunistic Infections/DIAGNOSIS/*MORTALITY  Great
       Britain/EPIDEMIOLOGY  Health Resources/*UTILIZATION  Human  Patient Care
       Team/*UTILIZATION  Pneumonia, Pneumocystis carinii/DIAGNOSIS/*MORTALITY
       Regression Analysis  Retrospective Studies  Survival Rate  MEETING
       ABSTRACT

       SOURCE: National Library of Medicine.  NOTICE: This material may be
       protected by Copyright Law (Title 17, U.S.Code).

