|
Intensive care unit prognostic scoring systems to predict death: a cost-effectiveness analysis |
Glance L G, Osler T, Shinozaki T |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology The use of intensive care unit (ICU) prognostic scoring systems to predict death.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised patients in an ICU, with the exception of cardiac surgical patients.
Setting The study setting was hospital. The economic study was carried out in the USA.
Dates to which data relate The effectiveness estimates were derived from data collected between January 1988 and April 1997, from studies published in 1994 and 1995. The cost data were derived from the database in a study published in 1993. The price year was not reported.
Source of effectiveness data The effectiveness data were derived from a literature review.
Modelling A Markov decision tree analysis was used to examine the cost-effectiveness of the two clinical strategies.
Outcomes assessed in the review The review assessed the sensitivity and specificity of the scoring system, decision point and ICU mortality rate.
Study designs and other criteria for inclusion in the review Sources searched to identify primary studies Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data Summary statistics from the database and individual studies were used.
Number of primary studies included Two primary studies were included.
Methods of combining primary studies Investigation of differences between primary studies Results of the review The specificity and sensitivity of the scoring system were 99.6 and 16.6%, respectively. The decision to withdraw life-sustaining care was made on ICU day three. The ICU mortality rate was 12.8%.
Estimates of effectiveness and key assumptions It was assumed that death resulted from the withdrawal of care, and that survival was until hospital discharge.
Measure of benefits used in the economic analysis The measure of benefits was the number of deaths prevented.
Direct costs The direct costs were not discounted due to the short time horizon of the study, i.e. less than one year. The quantities and costs were not reported separately. The direct costs were calculated using the Therapeutic Intervention Scoring System (TISS) scores from a database with $300 as the TISS point. The quantity/cost boundary adopted was that of the hospital. The price year was not reported.
Statistical analysis of costs Indirect Costs Indirect costs were not included.
Sensitivity analysis One-way sensitivity analyses were conducted on the mortality rate, specificity, the decision point to withdraw life-sustaining care, and the production cost per TISS point.
Estimated benefits used in the economic analysis The survival rates were reported as percentages only: 86.85% when a scoring system was used and 87.2% when one was not used.
Cost results The total costs, on the basis of intervention, were not given. The costs per patient were $29,200 and $30,100, with and without the use of a scoring system respectively.
Synthesis of costs and benefits The costs per survivor were $33,600 when a scoring system was used and $34,500 when one was not used. The cost per death prevented by not using a scoring system was $263,700; this was reduced to $21,700 by decreasing the specificity to 0.95.. The cost per death prevented increased rapidly as the hospital death rate increased. The cost per death prevented was a linear function of the production cost per TISS point.
Authors' conclusions Unless daily mortality risk estimates based on APACHE III can be shown to retain the same level of predictive power in ICUs outside the development database, it is unlikely that the incremental cost-effectiveness gained by using them as the basis to withdraw care is sufficient to justify their use in this manner.
CRD COMMENTARY - Selection of comparators A justification was given for the comparator used, namely that it was a current clinical strategy. You, as a user of the database, should decide if these health technologies are relevant to your setting.
Validity of estimate of measure of effectiveness The authors derived effectiveness estimates from a large database, although they used different published sources. The validity of the results was enhanced by sensitivity analyses to account for variability in the estimates.
Validity of estimate of measure of benefit The estimation of benefits was obtained directly from the effectiveness analysis.
Validity of estimate of costs Although only an estimate based on the TISS score was used, the validity of cost results was enhanced by appropriate sensitivity analyses. However, the quantities and prices were not reported separately, which limited the generalisability of the results. The price year was not reported which would make reflation exercises in other settings problematic.
Other issues The authors made appropriate comparisons of their findings with those from other studies, and addressed the issue of generalisability to other settings. The authors did not appear to present their results selectively. There were limited details of the baseline characteristics of the ICU patients enrolled in the study; this varied with ICU. The authors identified several limitations: the effectiveness estimates were obtained from data collected over 9 years, which may no longer reflect current levels of discharge and death; it was assumed that the probability distributions for the survivor and non-survivor cohorts in the database could be generalised to other ICUs with different mortality rates and patient mixes; and it was assumed that the probability distributions would not be affected in the scoring system arm after withdrawing care in patients with a mortality risk estimate greater than 90%. The authors did not explore the effect of changing the sensitivity of the scoring system on its effectiveness. They justified this on the basis that preventing false-positives is paramount and any export of the scoring system would probably reduce overall accuracy.
Implications of the study The authors state that "unless daily mortality risk estimates based on APACHE III can be shown to retain the same level of predictive power in ICUs outside the development database, it is unlikely that the incremental cost-effectiveness gained by using them as the basis to withdraw care is sufficient to justify their use in this manner." They recommend that the analysis is expanded to look at the effect of ICU mortality rate on the performance of the scoring system.
Bibliographic details Glance L G, Osler T, Shinozaki T. Intensive care unit prognostic scoring systems to predict death: a cost-effectiveness analysis. Critical Care Medicine 1998; 26(11): 1842-1849 Indexing Status Subject indexing assigned by NLM MeSH APACHE; Adult; Cost-Benefit Analysis; Death; Decision Support Techniques; Euthanasia, Passive; Health Services Research; Hospital Costs /statistics & Hospital Mortality; Humans; Inpatients /classification; Intensive Care Units /economics /utilization; Markov Chains; Medical Futility; Patient Selection; Prognosis; Prospective Studies; Sensitivity and Specificity; Survivors /statistics & Vermont; Withholding Treatment; numerical data; numerical data AccessionNumber 21998001820 Date bibliographic record published 28/02/2002 Date abstract record published 28/02/2002 |
|
|
|