EFFECT OF VARYING THRESHOLD AND SELECTIVE VERSUS UNIVERSAL

Document Sample
EFFECT OF VARYING THRESHOLD AND SELECTIVE VERSUS UNIVERSAL Powered By Docstoc
					                                                                                            Arch Iranian Med 2004; 7 (4): 267 – 271



                                                        Original Article


EFFECT OF VARYING THRESHOLD AND SELECTIVE VERSUS
 UNIVERSAL STRATEGIES ON THE COST IN GESTATIONAL
                DIABETES MELLITUS
                                                 •
                    Bagher Larijani MD , Arash Hossein-Nezhad MD, Ali-Reza Vassigh MD


            Background – The prevalence of gestational diabetes mellitus (GDM) is increasing globally,
        and the major determinants of screening programs are cost-benefit and prevalence in the target
        population. We aimed at assessing the screening method best-suited for a selected population in
        Tehran.
            Methods – A total number of 2,416 pregnant women were classified into high-, intermediate-,
        and low-risk groups, according to the American Diabetic Association (ADA) criteria. They were
        then screened for GDM at Tehran’s teaching hospitals using the two-step approach with thresholds
        of 130 mg/dL and 140 mg/dL, and the diagnostic criteria advocated by Carpenter and Coustan.
            Results – The prevalence of GDM in our sample was 4.7%. Switching from the 130 mg/dL to
        the 140 mg/dL threshold, decreased case-detection sensitivity by 12% (to 88%). With this
        approach, however, the cost of screening per pregnancy dropped from US $3.80 to US $3.20 (-
        15.6%), and the cost per detected case of GDM, from US $80.56 to US $77.43 (–3.9%).
            Conclusion – We recommend a universal screening for populations in whom there is a
        significant prevalence of GDM and variable health-care coverage, leading to a significant
        proportion of cases being missed, but where the cost of universal screening is markedly lower
        than that in more developed economies.

       Archives of Iranian Medicine, Volume 7, Number 4, 2004: 267 – 271.

       Keywords: Cost analysis • developing country • diagnosis • gestational diabetes mellitus •
        screening



                        Introduction                                    report of the ADA Expert Committee on Diagnosis
                                                                        and Classification of Diabetes Mellitus in that year,


    T     here is a general consensus that the
          prevalence of gestational diabetes
          mellitus (GDM) is increasing globally.1
In spite of more than 30 years of research,
                                                                        however, switched its recommendation to
                                                                        selective, or risk factor-based screening, a decision
                                                                        endorsed by the Fourth International Workshop-
                                                                        Conference on Gestational Diabetes Mellitus in
however, there is no consensus as yet regarding the                     1998.3 – 8
need for screening, diagnostic criteria, treatment,                        The approach adopted by most obstetricians in
or even the validity of GDM as a meaningful                             the USA is that advocated by the National Diabetes
diagnosis.2                                                             Data Group (NDDG) and the American College of
   Until 1997, the position adopted by the                              Obstetricians and Gynecologists (ACOG).9 – 12 The
American Diabetic Association (ADA) and the                             ACOG, being reluctant to recommend universal
First Three International Workshop-Conferences                          screening, instead endorsed screening protocols
on GDM, was to endorse universal screening. The                         based on a variety of risk factors.12, 13
                                                                           Universal screening or diagnostic testing is still
Authors affiliation: Endocrinology and Metabolism Research Center,      recommended for women in ethnic groups in
Tehran University of Medical Sciences, Tehran 14114, Iran.
•Corresponding author and reprints: Bagher Larijani, MD,                which the prevalence of carbohydrate intolerance
Endocrinology and Metabolism Research Center, Shariati Hospital,        during pregnancy, and the likelihood of developing
North Kargar Avenue, Tehran 14114, Iran. Fax: +98-21- 8029399,
E-mail: emrc@sina.tums.ac.ir.
                                                                        diabetes later in life, are relatively high. This


                                                                Archives of Iranian Medicine, Volume 7, Number 4, October 2004 267
                                              Effect of varying threshold and selective versus universal strategies on the cost in GDM



would include people of Hispanic, African, Native                       and Coustan.1, 16, 17
American, and South or East Asian, Polynesian,                              All borderline or suspicious test results were
and Australian aboriginal descent, especially if                        repeated. Every blood sample was centrifuged for
they lead a ‘western’ lifestyle or live in an urban                     a maximum of two hours after it was drawn. The
setting.1                                                               supernatant plasma were kept in a cold state and
    In developing countries, Iran being an example,                     transported to Shariati Hospital Laboratory where
limited budget resources preclude the delivery of                       they were tested. Plasma glucose was measured by
technology, or manpower-intensive care to the                           the glucose oxidase method, using a Hitachi 704
entire population. Cost-effectiveness and cost-                         autoanalyer. After universal testing was completed,
benefit calculations are, therefore, intrinsic to the                   samples were re-tested as if selective or risk factor-
design and delivery of health-care in these                             based screening had been used. We estimated the
countries. GDM in Iran has only been properly                           direct costs incurred by the hospital laboratory in
studied in the capital, Tehran, where the                               the screening and diagnosis of GDM in
prevalence is around 5%.14,15                                           consecutive pregnant women with historical and
    As the first step in determining the cost-                          clinical risk factors by proration of component unit
effectiveness of GDM screening in Iran, we                              costs. The results obtained with the two screening
studied the cost of different screening approaches                      approaches were analyzed and compared. The cost
and thresholds.                                                         of universal and selective screening methods were
                                                                        calculated based on both public and private sector
               Patients and Methods                                     tariffs, as was the cost of each portion of the GCT
                                                                        and OGTT tests, assuming use of standard material
    A total number of 2,416 pregnant women were                         and services, under the heading of calculated
enrolled in an observational study of GDM at four                       tariffs. The estimated unit cost in 2002 was $0.35
university teaching hospitals in Tehran. The risk                       for serum glucose determination, $0.65 for test
factors examined were known diabetes in first-                          solution for the GCT, $1.31 for test solution for the
degree relatives, a history of poor obstetric                           OGTT, and $0.32 for each phlebotomy. Finally,
outcome (spontaneous abortion, neonatal death,                          for a more realistic calculation of costs, the mean
intrauterine fetal death, anomaly, preterm labor),                      value of the public sector and private sector were
polyhydramnios, a history of having a macrosomic                        calculated and used as the basis of cost analysis
child, glycosuria, maternal age ≥ 35, and obesity.                      studies (Table 1). The cost of each screening
High-risk patients were screened during their first                     method was calculated as follows: cost per
antenatal visit. If they were found not to have                         pregnant woman = [(number of performed GCT*
GDM at an initial screening, they were retested                         average cost of GCT) + (number of performed
between the 24th and 28th weeks of their gestation.                     OGTT* average cost of OGTT)] /total pregnant
Remaining women (without risk factors) were also                        women enrolled; and cost per GDM case detected
screened for GDM between the 24th and 28th weeks                        = [(number of performed GCT* average cost of
of their pregnancy. All women with impaired                             GCT) + (number of performed OGTT* average
glucose tolerance and/or symptoms suggestive of                         cost of OGTT)]/total GDM patients detected. The
hyperglycemia were followed up and re-tested                            figures obtained were converted into US dollars to
between the 32nd and 36th weeks of pregnancy.                           allow comparison with studies done elsewhere
The 50 g glucose challenge test (GCT) with a                            (Tables 2 and 3).
threshold of 130 mg/dL was used as screening test,
and the 100 g oral glucose tolerance test (OGTT)                                                   Results
as the diagnostic test. Two or more abnormal
OGTT readings were considered diagnostic of                                A total number of 2,416 women underwent
GDM, based on the criteria proposed by Carpenter                        universal screening for GDM using the two-step

 Table 1. Average cost of diagnostic and screening tests for gestational diabetes mellitus in Tehran, Iran.
              Public sector tariff             Private sector tariff              Calculated cost                   Average
 Test
              Rials          US$              Rials              US$              Rials       US$             Rials         US$
 GCT         8,500           1.06            11,000              1.38            11,150       1.40           10,220         1.27
 OGTT        13,100          1.63            24,500              3.06            33,500       4.19           23,700         1.96
 GCT = 50-gram 1-hour glucose challenge test; OGTT = 100 g 3-hour oral glucose tolerance test; Exchange rate = 8,000 Rials = US $1,00;
 Calculated cost = cumulative cost of each portion of the test, assuming standard material and services.



268 Archives of Iranian Medicine, Volume 7, Number 4, October 2004
                                                                                           B. Larijani, A. Hossein-Nezhad, A.R. Vassigh


Table 2. Number of GCT and OGTT performed                                  approach missed 14% of the cases but reduced the
with different screening strategies.                                       number of GCTs by 21%, and OGTTs by 21.72%.
                                                           GDM             Universal screening with a 140 mg/dL threshold
Approach          Threshold        GCT       OGTT          cases
                                                          detected         reduced the number of GCTs by 43.5%, but also
                  130 mg/dL        4,499        1,160        114           missed 23% of GDM cases.
Universal
                  140 mg/dL        4,443        698          100              Analysis      of    the    direct     cost    of
                  130 mg/dL        3,025        908          98            diagnostic/screening tests for the ‘universal 130’
Selective
                  140 mg/dL        3,002        655          88
GCT = 50-gram 1-hour glucose challenge test; OGTT = 100-gram 3-
                                                                           approach yields a cost of US $3.80 (30,410 Rials)
hour oral glucose tolerance test; Threshold = 1-hour blood glucose         per patient screened and US $80.56 (644,488
value on GCT at or above which patients are referred for diagnostic        Rials) per case of GDM detected.               The
(OGTT) testing.
                                                                           corresponding values for the ‘universal 140’
approach with a 130 mg/dL threshold, based on the                          approach are US $3.20 (25,641 Rials) and US
Carpenter and Coustan criteria. Of these, 114                              $77.43 (619,500 Rials), respectively, but, as
(4.7%) were diagnosed with GDM. Screening was                              mentioned already, with 12% of cases being
performed before the 24th week of pregnancy in                             missed. Therefore, a 15.6% reduction in per capita
1,209 women, of whom 392 had an abnormal GCT                               screening cost and a 3.9% reduction in cost per
and, therefore, underwent a diagnostic OGTT.                               case detected may be achieved with a 12%
GDM was eventually diagnosed in 28 of the 392                              decrease in screening sensitivity (Tables 2 and 3).
patients with initial abnormal GCT. Between the                            The ‘selective 130’ and ‘selective 140’ methods
24th and 28th weeks of pregnancy, 2,388 women                              yield per capita screening costs of US $2.71
were screened: 517 went on to have a diagnostic                            (21,703 Rials) and US $2.39 (19,124 Rials),
OGTT and 44 were eventually diagnosed with                                 respectively, and cost per case detected of US
GDM. At 29th week of gestation, 906 women had                              $66.88 (535,052 Rials), and US $65.53 (525,044
either symptoms of hyperglycemia or a single                               Rials), respectively. The sensitivity and required
abnormal OGTT reading, but were otherwise                                  expenditure for each approach are shown in Table
normal. They were followed up and re-tested; 251                           3. Demographic and reproductive characteristics of
had an abnormal GCT during the second test.                                GDM and normal pregnant women are shown in
Forty-two were eventually diagnosed with GDM.                              Table 4.
Table 2 shows the total number of GCT and OGTT
performed with each screening protocol.                                                             Discussion
    Therefore, out of 114 patients diagnosed with
GDM, 24.56% were diagnosed before the 24th                                     Many authors have evaluated the direct costs of
week of pregnancy, 38.59% were diagnosed                                   different screening approaches for GDM. Several
between the 24th and 28th weeks of pregnancy, and                          investigators have examined the efficiency of
36.8% after the 29th week of pregnancy. Using a                            historical risk factors at narrowing the group to be
threshold of 140 mg/dL, the number of OGTT                                 screened. They have found these risk factors only
performed would have decreased by 39.8%, at the                            in roughly half of the women known to have
expense, however, of 12% of GDM cases being                                GDM.18, 19 Reed20 showed that 50% of all referrals
missed.                                                                    had risk factors for GDM; 48% of these had one
    Using selective screening with a 130 mg/dL                             risk factor and were older than 25 years; 85% of
threshold, 748 women were assessed before the                              patients with GDM belonged to this group. The per
24th week of pregnancy, 1,689 women between the                            capita cost of this approach was US $6.83,
24th and 28th weeks of pregnancy, and 698 women                            compared with US $24.40 for universal screening.
during the 32nd week of pregnancy. Overall, 98                             If only women over the age of 25 had been
women with GDM were eventually screened. This                              screened, the per capita cost would have risen to
 Table 3. Calculated sensitivity and cost of different screening strategies.
                                                                      Cost per pregnant woman          Cost per GDM case detected
 Approach                 Threshold             Sensitivity
                                                                         Rials          US$               Rials            US$
                          130 mg/dL                 100                 30,410          3.80             644,488           80.56
 Universal
                          140 mg/dL                 88                  25,641          3.20             619,500           77.43
                          130 mg/dL                 86                  21,703          2.71             535,052           66.88
 Selective
                          140 mg/dL                 77                  19,124          2.39             525,044           65.63
 Threshold = 1-hour blood glucose value on GCT at or above which patients are referred for diagnostic (OGTT) testing; Exchange rate = 8,000
 Rials = US $1.00; Sensitivity = true positives/(true positives + false negatives).



                                                                 Archives of Iranian Medicine, Volume 7, Number 4, October 2004 269
                                                   Effect of varying threshold and selective versus universal strategies on the cost in GDM


Table    4.    Demographic and reproductive                                in his study.24
characteristics of GDM and normal pregnant                                     In spite of the wealth of data, no clear picture
women.                                                                     emerges from the most appropriate screening
                      †
                       Normal               GDM
Characteristic
                     pregnancies           patients
                                                          pValue*          strategy for GDM. The variety of thresholds and
Maternal age         24.92 ± 5.31        29.09±6.13        < 0.000         selection protocols used, as well as the wide range
BMI (kg/m2)          24.92 ± 2.09        27.43±4.33        < 0.000         of GDM prevalence in the populations studied,
Parity                0.9 ± 1.34          1.79±2.09        < 0.01          make direct comparison very difficult. The present
FH**                    10.3%               33.3%          < 0.000         lack of consistency regarding different practice
Macrosomia**             3.4%               25.4%          < 0.000
Abortion**               9.4%               24.5%          < 0.01
                                                                           guidelines/recommendations         of   the     major
Delivery < 37                                                              professional/advocacy bodies in this field has not
                         3.2%               8.7%            < 0.01
weeks**                                                                    helped resolve the controversy. In addition, few
Glycosuria               2.2%              34.2%           < 0.000         studies have measured the cost-effectiveness or
Hydramnios               1.5%               5.2%            0.06           cost-benefit of screening for GDM, especially with
At least one
risk factors
                        40.6%              68.4%            < 0.01         pregnancy outcomes as a major end-point.25 – 27
††
  Selective                                                                The purchasing power of the client further makes
                         68%               85.9%            < 0.01
screening                                                                  the interpretation of simple cost saving difficult.
*Two-sided Student’s t-test in mean of age and BMI; Fisher’s exact             In this study, the prevalence of GDM was 4.7%,
test and Chi-square in other values; **Family history in first-degree
relative; † Women with a normal screening test (GCT); †† Women with        which is higher than many of the populations in
at least one risk factor or older than 24 years.                           whom cost analysis studies have been carried out.
                                                                               The cost of GDM screening strategies is a
US $7.34 and 24% of GDM cases would have been                              function of the testing protocol used, patient
missed.20, 21 In our study, 40.6% of all referrals had                     selection criteria, the blood glucose value at or
historical risk factors and 68% of all referrals had                       above which OGTT is performed, the demographic
historical, clinical, and reproductive risk factors or                     and fertility characteristics of the population under
were older than 24 years (Table 4). Callonge                               consideration, and the baseline prevalence of GDM
showed that universal screening with a 130 mg/dL                           in this population.
threshold has a sensitivity of nearly 100%, but                                Studies that advocate a selective screening
costs US $905 per case of GDM detected, and                                approach tend to be conducted in populations with
calculated the selective screening showing the cost                        a low prevalence of GDM, hence the low number
per case of GDM reducing to US $486 (46%                                   of missed cases—the sensitivity of a test
reduction), while missing only 8 cases of GDM per                          (screening) does depend on the baseline prevalence
1000.22                                                                    of positive (GDM) cases in the population.
   Coustan et al23 assessed the effect of different                            In our population and any other populations
screening strategies on cost and sensitivity.                              with similar demographic and socioeconomic
Universal screening with a 130 mg/dL threshold                             conditions, where the baseline prevalence of GDM
was 100% sensitive and cost US $249.00 per case                            is moderate to high and a significant number of
of GDM. Corresponding values for universal                                 cases are missed with selective screening,
screening with a 140 mg/dL threshold were 90%                              universal screening is the better option. Finally,
and US $222.00 (10.8% reduction), whereas                                  risk factor screening is not the recommended
limiting screening to women over the age of 25,                            method of screening for GDM in our antenatal
with a 140mg/dL threshold reduced the cost per                             population when the GDM prevalence, number of
case of GDM to US $192.00 (22.8% reduction)                                mothers with identified risk factors, cost
and sensitivity to 85%.23 In this study, a 15.6%                           effectiveness, and the risk of missing cases are
reduction in per capita screening cost and a 3.9%                          taken into consideration.
reduction in cost per case detected may be
achieved with a 12% decrease in screening                                                         References
sensitivity for using 140 mg/dL threshold instead
of 130 mg/dL threshold in universal screening.                             1    Metzger     BE,     Coustan    DR.    Summary     and
   Kitzmiller reported that risk factor-based                                   recommendations of the Fourth International Workshop-
screening leads to a 22% overall reduction in the                               Conference on Gestational Diabetes Mellitus. The
number of screening tests performed and has a                                   Organizing Committee. Diabetes Care. 1998; 21 (suppl
                                                                                2): B161 – 167.
sensitivity of 95% with a 130mg/dL threshold.                              2     Jarrett RJ. Gestational diabetes: a non-entity? BMJ.
The cost per case of GDM detected was US $215                                   1993; 306: 37 – 38.



270 Archives of Iranian Medicine, Volume 7, Number 4, October 2004
                                                                                    B. Larijani, A. Hossein-Nezhad, A.R. Vassigh


3    Expert Committee on the Diagnosis and Classification of             of gestational diabetes mellitus in Tehran. Proceedings of
     Diabetes Mellitus. Report of the Expert Committee on the            the Third International Congress on Endocrine Disorders.
     Diagnosis and Classification of Diabetes Mellitus.                  Tehran; 1995: 4 – 8.
     Diabetes Care. 1997; 20: 1183 – 1197.                          15   Larijani B, Azizi F, Hossein-Nezhad A, et al. Prevalence
4    Danilenko-Dixon DR, van Winter JT, Nelson RL,                       of gestational diabetes mellitus in pregnant women
     Ogburn PL Jr. Universal versus selective gestational                referred to hospitals affiliated to Tehran University of
     diabetes screening: application of the 1997 American                Medical Sciences. Iranian J Endocrinol Metab 1999; 1:
     Diabetes Association recommendations. Am J Obstet                   125 – 133.
     Gynecol. 1999; 181: 798 – 802.                                 16   Sweeney AT, Brown FM. Gestational diabetes mellitus.
5    American Diabetes Association. Gestational diabetes                 Clin Lab Med. 2001; 21: 173 – 191.
     mellitus: American Diabetes Association position               17   American Diabetes Association 60th Scientific Sessions.
     statement. Diabetes Care. 1986; 4: 430 – 431.                       Diabetes and pregnancy. Diabetes Care. 2000; 23: 1699
6    Frienkel N, Josimovich J. Conference Planning                       – 1792.
     Committee: American Diabetes Association Workshop-             18   Lavin JP, Barden TP, Miodovnik M. Clinical experience
     Conference on Gestational Diabetes: summary and                     with a screening program for gestational diabetes. Am J
     recommendations. Diabetes Care. 1980; 3: 499 – 501.                 Obstet Gynecol. 1981; 141: 491 – 494.
7    Frienkel N. Summary and recommendations of the                 19   Marquette GP, Klein VR, Niebyl JR. Efficacy of
     Second      International   Workshop-Conference        on           screening for gestational diabetes. Am J Perinatol. 1985;
     Gestational Diabetes Mellitus. Diabetes. 1985; 34 (suppl            2: 7 – 9.
     2): 123 – 126.                                                 20   Reed BD. Screening for gestational diabetes: analysis by
8    Metzger BE. Summary and recommendations of the                      screening criteria. J Fam Pract. 1984; 19: 751 – 755.
     Third Workshop-Conference on Gestational Diabetes              21    Massion C, O'Connor PJ, Gorab R, Crabtree BF,
     Mellitus. Diabetes. 1991; 40 (suppl 2): 197 – 201.                  Nakamura RM, Coulehan JL. Screening for gestational
9    Lavin JP Jr, Lavin B, O'Donnell N. A comparison of                  diabetes in a high-risk population. J Fam Pract. 1987;
     costs associated with screening for gestational diabetes            25: 568 – 575.
     with two-tiered and one-tiered testing protocols. Am J         22   Callonge N. Commentary on screening for gestational
     Obstet Gynecol. 2001; 184: 363 – 367.                               diabetes. J Fam Pract. 1987; 25: 575 – 576.
10   Landon MB, Gabbe SG, Sachs S. Management of                    23   Coustan DR, Nelson C, Carpenter MW, Carr SR,
     diabetes mellitus in pregnancy: a survey of obstetricians           Rotondo L, Widness JA. Maternal age and screening for
     and maternal-fetal medicine specialists. Obstet Gynecol.            gestational diabetes: a population-based study. Obstet
     1990; 75: 635 – 640.                                                Gynecol. 1989; 73: 557 – 561.
11   National Diabetes Data Group. Classification and               24   Kitzmiller JL. Cost analysis of diagnosis and treatment of
     diagnosis of diabetes mellitus and other categories of              gestational diabetes mellitus. Clin Obstet Gynecol. 2000;
     glucose intolerance. National Diabetes Data Group.                  45: 140 – 153.
     Diabetes. 1979; 28: 1039 – 1057.                               25   Everett WD. Screening for gestational diabetes: an
12   American College of Obstetricians and Gynecologists.                analysis of health benefits and costs. Am J Prevent Med.
     Management of diabetes mellitus in pregnancy.                       1989; 5: 38 – 43.
     Washington: The College; 1994. ACOG Technical                  26   Kitzmiller JL, Elixhauser A, Carr S, et al. Assessment of
     Bulletin No. 200.                                                   costs and benefits of management of gestational diabetes
13   American College of Obstetricians and Gynecologists.                mellitus. Diabetes Care. 1998; 21 (suppl 2): B123 – 137.
     Management of diabetes mellitus in pregnancy.                  27    Langer O, Rodriguez DA, Xenakis EM, McFarland MB,
     Washington: The College; 1986. ACOG Technical                       Berkus MD, Arrendondo F. Intensified versus
     Bulletin No.92.                                                     conventional management of gestational diabetes. Am J
14   Larijani B, Bastanhagh MH, Pajouhi M, et al. Prevalence             Obstet Gynecol. 1994; 170: 1036 – 1046.




                                                            Archives of Iranian Medicine, Volume 7, Number 4, October 2004 271

				
DOCUMENT INFO