|
Yield and cost of individual common diagnostic tests in new primary care outpatients in Japan |
Takemura Y, Ishida H, Inoue Y, Beck J R |
|
|
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 A test package, designated the Essential Laboratory Test 2 (ELT-2), was under investigation for the initial clinical diagnosis and identification of occult disease. The ELT-2 extensive test panel comprised basic tests in the ELT-1, specific tests and some optional tests. The basic tests in the ELT-1 are dipstick urinalysis, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), haemoglobin, red blood cell (RBC) indices, white blood cell count (WBC), total protein, albumin and albumin/globulin ratio (A/G). The specific tests included in the ELT-2 are urine sediment, sialic acid, platelet count, leukocyte differential count (LDC), serum protein fraction profile, total cholesterol, glucose, aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LD), alkaline phosphatase (ALP), gamma-glutamyltransferase (GGT), cholinesterase, serum urea nitrogen, creatinine and uric acid. The optional tests included faecal occult blood, chest X-ray, abdominal X-ray and electrocardiogram (ECG). The comparators were each of these components of the ELT-2 and also the ELT-1.
Type of intervention Diagnosis and primary prevention.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised new outpatients presenting between the dates of the study with defined symptoms, and who were seen by physicians. Those patients referred by physicians in other medical facilities with test results and/or tentative clinical diagnoses were not included.
Setting The setting was primary care. The economic study was carried out in Japan.
Dates to which data relate The effectiveness data were collected between June 1991 and March 1997. The dates for when the resource data were collected were not reported. A price year was not given.
Source of effectiveness data The effectiveness data were derived from a single study.
Link between effectiveness and cost data The costing was carried out retrospectively on a different sample of patients to that used in the effectiveness study.
Study sample The authors have published two related analyses based on the same clinical study (see Other Publications of Related Interest). The sample included all patients presenting at the setting during the dates of the study, who were eligible according to the study population. No further exclusions were made. The authors did not report that power calculations were carried out to estimate the impact of chance on the results. The study included 540 patients (250 males) aged from 9 to 83 years. The sample was appropriate for the clinical question, as it included patients with defined symptoms who might have asymptomatic diseases that could be diagnosed via the panel tests.
Study design The basis of the analysis was a diagnostic study. The study was conducted at two centres, the Comprehensive Medicines Clinic, National Defense Medical College Tokorozawa, Japan, and its affiliated hospital. The analysis was based on the immediate results of diagnostic tests, meaning that there was no requirement for patient follow-up.
Analysis of effectiveness The analysis of the clinical study was based on the number of tests yielding a useful result (UR) and the number of tests contributing to a case finding (i.e. discovery of occult disease). A UR was defined as a test result contributing to a change in a physician's behaviour or decision-making between the Tentative Initial Diagnoses (TID) and the Initial Clinical Diagnosis (ICD). The TID was made by the primary physician based on the patient's history and physical examination alone. The ICD was made after integrating the results of diagnostic tests. One UR was assigned for each TID. Therefore, the primary outcomes were the number and proportion of URs and case findings. Since all patients received the ELT-2 panel tests there was only a single study group and no need for inter-group comparisons. The authors have assessed the sensitivity, specificity and positive predictive value of each test (Takemura et al., see Other Publications of Related Interest).
Effectiveness results The ELT-2 panel test yielded 398 URs and uncovered 261 occult diseases among 540 patients.
When optional test items were excluded, the contribution rates of individual tests to UR-generation ranged from 0.252 for CRP to 0.006 for uric acid. The contribution rates to case finding ranged from 0.184 for total cholesterol to 0.004 for serum urea nitrogen and creatinine.
The proportion of ELT-2 specific tests yielding a UR or contributing to a case finding was:
for urine sediment, UR generation 0.037, case finding 0.048;
for sialic acid, UR generation 0.219, case finding 0.009;
for platelet count, UR generation 0.035, case finding 0.019;
for LDC, UR generation 0.231, case finding 0.010;
for serum protein fraction profile, UR generation 0.192, case finding 0.064;
for total cholesterol, UR generation 0.076, case finding 0.184;
for glucose, UR generation 0.019, case finding 0.031;
for AST, UR generation 0.089, case finding 0.041;
for ALT, UR generation 0.086, case finding 0.078;
for LD, UR generation 0.062, case finding 0.016;
for ALP, UR generation 0.035, case finding 0.012;
for GGT, UR generation 0.038, case finding 0.046;
for cholinesterase, UR generation 0.063, case finding 0.088;
for serum urea nitrogen, UR generation 0.014, case finding 0.004;
for creatinine, UR generation 0.008, case finding 0.004;
for uric acid, UR generation 0.006, case finding 0.032.
The ELT-2-specific test items contributed more frequently to the identification of occult disease than the basic tests in the ELT-1 panel.
Total cholesterol frequently contributed to UR-generation and case finding in most disease categories, whereas the contribution rates of other tests were highly dependent on the disease categories.
In summary, 1,592 tests of 12,646 tests performed in all patients contributed to either UR-generation or discovery of occult disease.
Clinical conclusions The authors did not draw conclusions for the individual tests, but did stress that the effectiveness of the tests varied and differed across disease categories.
Measure of benefits used in the economic analysis The summary measures of benefit used were the number of URs and the number of tests contributing to a case finding.
Direct costs The authors did not report a perspective for the costing analysis. Discounting was not required as the analysis was concerned with the immediate costs of carrying out the tests. The cost estimates were calculated by considering all actual expenditures required to obtain test results at the Kawasaki Medical School Hospital (Kurashiki, Japan). The cost estimates covered test reagents, consumptive materials required for operation of analysers, computers and printers, disposables for specimen collection and sample analysis, equipment amortisations, and personnel expenses for medical technologists. The authors made adjustment for multi-channel organisers. The unit costs were reported. The dates when the resources were collected and the price year were not reported. If the costs were collected for the full duration of the clinical trial, then adjustments would be required to reflate the costs to the chosen price year.
Statistical analysis of costs The costs were treated deterministically.
Indirect Costs Indirect costs were not estimated, and would not have been relevant if the perspective of a third-party payer or hospital was used.
Sensitivity analysis There was no report of sensitivity analyses being carried out.
Estimated benefits used in the economic analysis See the 'Effectiveness Results' section.
Cost results The figures presented represent the actual unit costs of the ELT-2-specific tests:
urine sediment, Y288.7;
sialic acid, Y118.7;
platelet count, Y56.8;
LDC, Y47.2;
serum protein fraction profile, Y185.0;
total cholesterol, Y29.4;
glucose, Y32.3;
AST, Y25.3;
ALT, Y25.3;
LD, Y23.1;
ALP, Y25.2;
GGT, Y24.1;
cholinesterase, Y28.1;
serum urea nitrogen, Y24.9;
creatinine, Y34.5; and
uric acid, Y27.6.
Synthesis of costs and benefits The cost per test yielding a UR varied from Y116 for CRP to Y7,731 for urine sediment. The cost per test contributing to a case finding ranged from Y153 for total cholesterol to Y12,739 for sialic acid.
The cost per test either yielding a UR or contributing to a case finding (named cost per effective test) varied from Y108 for total cholesterol to Y6,200 for chest X-ray.
For each ELT-2-specific test, the cost per effective test was:
urine sediment, Y3,394;
sialic acid, Y520;
platelet count, Y1,043;
LDC, Y196;
serum protein fraction profile, Y721;
total cholesterol, Y108;
Glucose, Y646;
AST, Y193;
ALT, Y154;
LD, Y294;
ALP, Y531;
GGT, Y284;
cholinesterase, Y185;
serum urea nitrogen, Y1,411;
creatinine, Y2,933; and
uric acid, Y728.
The cost per effective test differed depending on the TID categories.
Thus, the authors redesigned the test panels to have an acceptable effectiveness in the aggregate for the respective patient groups. Thus, the incremental cost-effectiveness ratios of these new panels against the ELT-1 baseline panel could be high for the cardiovascular and respiratory panel (Y1,249 and Y493 per additional effective test gained, respectively), whereas other redesigned panels could achieve cost-savings compared with the ELT-1 baseline panel (e.g. -$1,140 and -$107 per additional effective test gained for gastrointestinal and metabolic/endocrine diseases, respectively).
Authors' conclusions The authors concluded that analyses of cost-effectiveness enabled them to establish new test panels "which differ from the ELT test panel and give maximal effectiveness at an acceptable cost, for selected groups of patients".
CRD COMMENTARY - Selection of comparators The authors aimed to compare the components of the ELT-2 panel test. In doing so they also presented results for the ELT-1 panel test and results for optional tests. This choice was justified in part by guidance from the Japanese Society of Clinical Pathology concerning the use of panel tests, and in part a desire to be more selective in ordering tests for patients. Current practice at the time of the study was to use ELT-1 in all patients and to use ELT-2 where further information was required.
Validity of estimate of measure of effectiveness The analysis was based along diagnostic lines. However, an analysis of sensitivity and specificity was not undertaken, partly because this information was already known for some diseases, but mainly because this study was concerned with occult diseases. Therefore, the use of sensitivity and specificity type outcomes would have been inappropriate. The study sample comprised all patients presenting at the study setting within the dates of the study and was representative of the study population. The authors provided summary results for diseases detected and broke these down by disease categories. Not all of the results are presented in this abstract. The reader should view the original paper for the complete results.
Validity of estimate of measure of benefit The estimation of a summary measure of benefits was taken directly from the effectiveness analysis. The authors used the number of tests contributing to URs and case findings as their measure. This seems a natural choice given the nature of the effectiveness study.
Validity of estimate of costs The perspective for the economic analysis was not reported, thus it is not possible to assess whether all the relevant costs were included. The costing analysis seems to have focused on the cost of obtaining the test results from the perspective of the hospital or third-party payer, focusing on the unit costs of consumables, machines and labour. The authors did not use statistical or sensitivity analyses to explore the impact of uncertainty in cost estimates. This makes it difficult for the reader to assess the implications of generalising the results to their own setting, and does not provide any indication of the certainty surrounding the actual estimates used in the analysis. Whilst the unit cost of each test was reported, there was no breakdown indicating the composition of this estimate and the relative importance of the cost elements.
Other issues The authors set their work in the context of their own prior work very well, but did not draw wider comparisons. It is unclear whether any work has been carried out that might have provided an appropriate comparison. The issue of generalisability was considered, with the authors acknowledging that more patients would have been preferable and that cost data will not be identical across primary care settings. However, the ability to generalise the current study would have been greatly increased by statistical and sensitivity analyses. Extensive results were presented in the study, only some of which have been presented in this abstract. Readers should view the paper to understand the complete results. The conclusions drawn accurately reflect the scope of the study, with the authors setting out to assess the cost-effectiveness of elements of ELT-2 in order to allow cost-effective testing. They have achieved this in establishing new test panels. A number of limitations were presented. These included the use of very short-term cost and effectiveness data, and the failure to consider negative effects of testing (e.g. misdiagnosis).
Implications of the study The authors did not explicitly make any recommendations for policy or practice, although they did suggest that "evidence elicited from this study can guide optimal usage of common diagnostic tests". They further suggested that studies involving more patients and multiple institutions would make ideal topics for further work.
Source of funding Supported in part by grants from the Clinical Pathology Research Foundation of Japan, Pfizer Health Research Foundation, and the Kurozumi Medical Foundation.
Bibliographic details Takemura Y, Ishida H, Inoue Y, Beck J R. Yield and cost of individual common diagnostic tests in new primary care outpatients in Japan. Clinical Chemistry 2002; 48(1): 42-54 Other publications of related interest Takemura Y, Ishida H, Inoue Y, Beck J. Common diagnostic test panels for clinical evaluation of new primary care outpatients in Japan: a cost-effectiveness evaluation. Clinical Chemistry 1999;45:1752-61.
Takemura Y, Ishida H, Inoue Y, Kobayashi H, Beck J. Opportunistic discovery of occult disease by use of test panels in new, symptomatic primary care outpatients: yield and cost of case finding. Clinical Chemistry 2000;46:1091-8.
Takemura Y, Kobayashi H, Kugai N, Sekiguchi S. The results of the "Essential Laboratory Tests" applied to new outpatients - re-evaluation of diagnostic efficiencies of the test items. Rinsho Byori 1996;44;555-63.
Indexing Status Subject indexing assigned by NLM MeSH Adolescent; Adult; Aged; Aged, 80 and over; Ambulatory Care /economics; Blood Chemical Analysis /economics; Child; Clinical Laboratory Techniques /economics; Cost-Benefit Analysis; Feces /chemistry; Female; Humans; Japan; Male; Middle Aged; Primary Health Care /economics; Urinalysis /economics AccessionNumber 22002006229 Date bibliographic record published 30/04/2005 Date abstract record published 30/04/2005 |
|
|
|