Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print June 11, 2003, 10.1148/radiol.2282020915
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2282020915v1
228/2/515    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McMahon, P. M.
Right arrow Articles by Gazelle, G. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McMahon, P. M.
Right arrow Articles by Gazelle, G. S.
(Radiology 2003;228:515-522.)

Evidence-based Practice

Cost-Effectiveness of PET in the Diagnosis of Alzheimer Disease1

Pamela M. McMahon, BS, Sally S. Araki, SM, Eileen A. Sandberg, SM, Peter J. Neumann, ScD and G. Scott Gazelle, MD, MPH, PhD

1 From the Institute for Technology Assessment, Massachusetts General Hospital, Zero Emerson Place, Suite 2H, Boston, MA 02114 (P.M.M., G.S.G.); Doctoral Program in Health Policy, Harvard University, Cambridge, Mass (P.M.M., S.S.A., E.A.S.); Program on the Economic Evaluation of Medical Technology, Harvard Center for Risk Analysis, Cambridge, Mass (S.S.A., E.A.S., P.J.N.); and Harvard School of Public Health, Boston, Mass (P.J.N., G.S.G.). P.M.M. supported in part by the National Cancer Institute grant 1 R25 CA92203-01A1 and the National Library of Medicine grant 5 T-15 LM07092. P.M.M. and G.S.G. supported in part by the U.S. Department of the Army under DAMD grant 17-99-2-9001. Received July 24, 2002; revision requested September 18; revision received September 25; accepted October 25. Address correspondence to G.S.G. (e-mail: scott@mgh-ita.org).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 REFERENCES
 
PURPOSE: To evaluate the cost-effectiveness of positron emission tomography (PET) in the diagnosis of Alzheimer disease (AD) in community-dwelling patients with mild or moderate dementia who present to specialized AD centers.

MATERIALS AND METHODS: A decision-analytic model was used to compare costs and quality-adjusted life years (QALYs) associated with strategies involving single photon emission computed tomography (SPECT), dynamic susceptibility-weighted contrast material–enhanced magnetic resonance (MR) imaging, and PET as functional imaging adjuncts to the standard clinical work-up. Sensitivity analyses were performed to examine changes in test characteristics, health-related quality-of-life survey instruments, therapeutic effectiveness, and treatment rules.

RESULTS: The use of PET to confirm the results of the standard clinical work-up cost more but yielded fewer benefits than a strategy in which dynamic susceptibility-weighted contrast-enhanced MR imaging was substituted for the typically performed structural computed tomography. This relationship remained stable in scenarios in which standard diagnostic work-up accuracy, drug treatment effectiveness, and version of the Health Utilities Index were altered. Dynamic susceptibility-weighted contrast-enhanced MR imaging cost $598,800 per QALY gained (range, $74,400 to $1.9 million per QALY), compared with the cost of the standard diagnostic work-up. Treating all patients with dementia was the dominant imaging strategy, except when side effects in patients with non–AD-related dementia were modeled. In all scenarios, SPECT yielded fewer benefits than other strategies at a higher cost.

CONCLUSION: PET may have high diagnostic accuracy, but adding it to the standard diagnostic regimen at AD clinics would yield limited, if any, benefits at very high costs.

Index terms: Alzheimer disease, 10.83 • Brain, PET, 10.12163 • Cost-effectiveness • Dementia, 10.83 • Positron emission tomography (PET), 10.12163


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 REFERENCES
 
Studies (1,2) have concluded that fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET) may be useful and acceptable for evaluating dementia and confirming the diagnosis of Alzheimer disease (AD). In January 2002, the Medicare Coverage Advisory Committee voted against reimbursement for FDG PET for the diagnosis of AD-related dementia (3). The final Center for Medicare and Medicaid Services decision will influence the medical care of the growing population of elderly Americans. The annual number of new-onset (ie, incident) cases of AD may reach 959,000 in year 2050 (4), and the total number of cases in the United States is expected to increase to 8.6 million by year 2047 (5,6). Reimbursement for PET performed for the diagnosis of AD would have a pronounced effect on Medicare expenditures; 2001 Medicare reimbursements for other covered PET procedures exceeded $2,000 per examination (7,8). However, the costs of medical care for patients with AD also are high (9–11), and proper treatment could result in lower care costs and a better quality of life for many patients with AD (12,13).

Given the lack of clear evidence of improved patient outcomes after diagnosis with PET, it is perhaps not surprising that there is controversy regarding insurers’ decisions of whether to reimburse for this examination. Some argue that reimbursement would be premature (14) or that PET has limited usefulness and is mainly a research tool (15). Others maintain that use of PET will both improve the quality of care for patients and save money (16).

In previous analyses (17), we evaluated the cost-effectiveness of adding either single photon emission computed tomography (SPECT) or dynamic susceptibility-weighted contrast material–enhanced magnetic resonance (MR) imaging to the clinical work-up of patients with AD. Our decision-analytic model is a combination of a diagnostic model and a model encompassing treatment and care setting (13, 17). Our purpose in the current analysis was to evaluate the cost-effectiveness of PET in the diagnosis of AD in community-dwelling patients with mild or moderate dementia who present to specialized AD centers.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 REFERENCES
 
General Analysis
In our base-case cost-effectiveness analysis, we used the best point estimates for all inputs (eg, parameters and event probabilities) and followed the consensus recommendations outlined by the Panel on Cost-effectiveness in Health and Medicine (18); a brief summary of key definitions and recommendations is provided: Strategies that cost more but yield fewer benefits than an alternative strategy are considered to be strongly dominated (ie, dominated). Benefits are expressed as gains in quality-adjusted life years (QALYs), which are weights of between 0 (patient dead) and 1.0 (patient in perfect health) assigned to each year of life. Costs are expressed in 1999 dollars (adjusted for inflation by using the medical care component of the consumer price index, except where noted). Future costs and QALYs are discounted at 3% annually. The main outcome measure for each strategy or intervention is the incremental cost-effectiveness ratio, or ICER, where changes in resource use (compared with the next best strategy) are divided by the QALYs gained (compared with the next best strategy): ICER = {Delta} costs/{Delta} QALYS. The study was conducted from a societal perspective.

Model Structure and Strategies Compared
The model we used has been described in detail elsewhere (13,17); a summary of the structure and key assumptions follows: The target population comprises community-dwelling patients with mild or moderate dementia who present to a specialized AD center. Following diagnosis and treatment (if any), patients transit between disease stages and care settings and accrue costs and benefits for each cycle. See the Appendix for estimates of disease prevalence and transition probabilities. The comparator, or the status quo strategy, is the standard clinical work-up, which includes acquisition of detailed medical history, assessment of cognition and functional status, laboratory testing, and structural brain imaging with nonenhanced computed tomography (CT) (19,20).

In this study, we compared strategies involving the use of either computed SPECT, dynamic susceptibility-weighted contrast-enhanced MR imaging, or FDG PET as functional imaging additions to the standard clinical examination. Visual SPECT was not considered, because it was dominated in all scenarios in our previous study (17). Note that in the base case, PET or computed SPECT was performed (in a second visit) only in those patients who received a diagnosis of possible or probable AD on the basis of the standard examination results; patients with positive PET or computed SPECT results then received therapy. In contrast, dynamic susceptibility-weighted contrast-enhanced MR imaging is an integrated structural and functional examination that can be performed instead of structural CT and thus was performed in all patients in that strategy, and patients were treated according to the results of this imaging examination.

An additional strategy, that of treating all patients with dementia (ie, treat all dementia), was one in which only the costs for consultation to confirm the diagnosis of dementia were included (ie, no imaging examinations, including structural imaging, or laboratory tests were conducted) and all patients were treated with donepezil hydrochloride (Aricept; Pfizer, New York, NY).

For each scenario, 100,000 Monte Carlo trials were simulated; the mean costs and mean effectiveness of each strategy were calculated, and incremental cost-effectiveness ratios were determined, as described earlier herein. The time frame of the base-case analysis was 18 months on the basis of the findings of a previously described survey of clinical experts (13). All model inputs and data sources are listed in the Appendix. The parameter estimates that have changed since our previous analysis are described in the following text.

Parameter Estimates for the Base Case
As stated earlier herein, patients with severe AD were excluded from our hypothetical patient population on the grounds that donepezil hydrochloride is not indicated for the treatment of severe AD, and because imaging would not be necessary to confirm the diagnosis. For the test characteristics of the standard clinical examination, we estimated sensitivities of 0.70 and 0.80 for the detection of mild and moderate AD, respectively, and a specificity of 0.73 for the detection of AD. These estimates were based on the results of published studies in which sensitivity values were estimated according to stage of disease. Morris et al (21) reported the standard clinical examination to have a sensitivity of 0.70 in 10 patients with mild dementia (also confirmed with pathologic analysis). In patients with more advanced AD, clinical examinations are expected to have higher sensitivities.

The results of two studies provide bounds for our estimate (0.80) of the sensitivity for detection of moderate AD: the lower bound of 0.75 (12 of 16 cases) determined by Hoffman et al (2) and the upper bound of 0.94 (45 of 48 cases) determined by Wade et al (22). The latter study included patients with moderate and severe AD and mixed diagnoses and involved a comparison between the clinical diagnoses (including those made with CT for most patients) and the autopsy results. Also, in a study that included pathologic confirmation of the diagnosis for 58 patients, Mölsä et al (23) reported a clinical examination sensitivity of 0.71 (20 of 28 cases) and a specificity of 0.73 (22 of 30 patients). That study had a patient population with a prevalence of AD equal to 0.48, which is roughly similar to the prevalence of AD that has been observed at Massachusetts General Hospital and was used in our base-case analysis (56%). Note that our base-case estimates agree with American Academy of Neurology estimates for the clinical diagnosis of probable AD (sensitivity of 0.81, specificity of 0.70) (20).

Base-case estimates of FDG PET sensitivity (0.94) and specificity (0.72) were derived from the results of studies (1,24) in which AD was differentiated from other causes of dementia (and not absence of dementia, ie, normal controls) in patients seen at AD referral centers.

For the cost of PET, we used a 1999 resource use estimate of $1,671, which was determined by using the cost-accounting software (Transition Systems, subsidiary of Eclypsis, Delray Beach, Fla) used at our institution (described in detail elsewhere [25]); Medicare does not reimburse for PET performed for diagnosis of AD-related dementia. For the cost of computed SPECT ($2,175), we added the cost for computer-aided data manipulation ($92, Medicare reimbursement for CPT 4 [Current Procedural Terminology 4] code 76375) to our institution’s 1999 Medicare reimbursement for visual SPECT imaging ($2,083, includes costs for professional and technical components). We estimated the cost for MR imaging plus dynamic susceptibility-weighted contrast-enhanced MR imaging, a relatively new procedure, to be equal to the Medicare reimbursements for MR imaging with and without contrast material plus the costs for a computerized three-dimensional reconstruction of the image data (total cost, $1,444). The total cost of the standard examination was estimated to be $533 (see Appendix).

Health-related quality-of-life weights based on the Mark III version of the Health Utilities Index (HUI3) (26,27) for patients with and without AD were used for the base-case analysis. The HUI3 is more sensitive than the Mark II version of the Health Utilities Index (HUI2) for severe impairments such as those seen with AD and therefore may be more appropriate for assessing quality of life in this population (28). The preference scores from HUI3 are substantially lower than those from HUI2 (used in our previous analysis). The HUI3 weights for patients with AD were derived from existing data (28) that were stratified by care setting (community or nursing home) for this model. HUI3 weights for age-matched community-dwelling Canadians were derived from the literature (28) and used for patients without AD.

Although new drugs such as rivastigmine tartrate (Exelon; Novartis, Basel, Switzerland) and galantamine hydrobromide (Reminyl; Janssen Pharmaceutica, Titusville, NJ) have been approved for the treatment of AD since our previous analysis was conducted, they are cholinesterase inhibitors like donepezil hydrochloride. All of these agents have similar effects but different side effect profiles (30). We retained our use of donepezil hydrochloride for the base-case analysis because we have precise estimates of the risk ratios for transitions between disease states (13) associated with this agent and addressed the potential improvements in therapies, as well as the potential side effects, in sensitivity analyses. To our knowledge, few data on the effects (beneficial or detrimental) of donepezil hydrochloride in patients with non–AD-related dementia exist; thus, for the base-case analysis, we assumed neither benefits nor side effects in these patients.

Sensitivity Analyses
We postulated the existence of a hypothetical perfect examination—that is, one with a sensitivity and specificity of 1.0—as an upper bound of gains in effectiveness achieved with improvements in test characteristics. A sensitivity analysis was performed to determine the robustness of the results achieved by changing the version of the HUI.

In our previous analysis (17), the sensitivity and specificity of the standard examination were critical parameters. We examined a scenario with a more lenient treatment rule (ie, treat possible AD as well as probable AD), which involved the use of a higher estimate for sensitivity (0.93) but a lower estimate for specificity (0.48) of the standard clinical examination for AD. These estimates were based on the averaged results of four published studies (20).

Improvements in therapies were modeled by more favorable (as compared with donepezil hydrochloride treatment) risk ratios for the transitions between mild and moderate AD. Drugs X and Y were postulated (see Appendix); the associated costs and durations of effectiveness were assumed to be the same as those for donepezil hydrochloride. Longer (48-month) and shorter (6-month) time horizons were used to simulate different durations of effectiveness (ie, lengths of time the risk ratio is in effect). Donepezil hydrochloride is a well-tolerated drug with few side effects (31); thus, in our base-case analysis, we assumed no decrease in the quality of life for patients with non–AD-related dementia who were treated inappropriately as a result of false-positive examination results. We modeled scenarios that included a decrease of 0.05 QALY for patients with non–AD-related dementia who were treated with donepezil hydrochloride and, in a separate analysis, with the hypothetical drug X.

Analyses were performed to examine changes in the test characteristics of PET. Hoffman et al (2) reported a sensitivity estimate (0.93) similar to that in our base case but a lower specificity estimate (0.63). We examined scenarios with a sensitivity of 0.94 (unchanged from the base case) and specificities of 0.45 and 1.00 (the 95% CI in the Hoffman et al study), as calculated by Silverman et al (1,24). Estimates of sensitivity for detection of mild AD range from 0.81 to 0.97, and those for detection of moderate AD range from 0.89 to 0.99 (1); the upper bounds are essentially the same as those of the hypothetical perfect examination, so they were not examined.

Finally, we modeled a scenario in which PET was offered to all patients who received a diagnosis of "AD unlikely or excluded" on the basis of standard examination results. The patients who received a positive diagnosis based on the results of either examination (ie, standard examination or PET) were treated with donepezil hydrochloride.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 APPENDIX
 REFERENCES
 
The results of the base-case analysis are listed in Table 1 and illustrated in the Figure. FDG PET was dominated by (ie, cost more but yielded fewer benefits) the standard clinical examination. Dynamic susceptibility-weighted contrast-enhanced MR imaging cost more and provided more benefits than the standard examination, with an incremental cost-effectiveness ratio of $598,800 per QALY. Computed SPECT was dominated by dynamic susceptibility-weighted contrast-enhanced MR imaging. The strategy of treating all patients who had dementia with donepezil hydrochloride dominated all of the functional imaging strategies considered and had an incremental cost-effectiveness ratio of $141,200 per QALY, as compared with the standard examination.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Base-Case Results

 


View larger version (22K):
[in this window]
[in a new window]
[Download PPT slide]
  		With our base-case assumptions, the addition of either PET or computed SPECT (cSPECT) to the standard clinical examination results in a strategy that costs more but yields fewer gains in QALYs. Dynamic susceptibility-weighted contrast-enhanced (ie, functional) MR imaging (DSC-MRI), as compared with the standard examination, yields gains in effectiveness at an incremental cost-effectiveness ratio of $598,800 per QALY. When no side effects in the patients with non-AD-related dementia who are treated with donepezil hydrochloride are included in the model, the strategy of treating all patients with dementia dominates all the imaging strategies, with an incremental cost-effectiveness ratio of $141,200 per QALY.

 
Note that our estimate of the cost of PET ($1,671) was based on resource use (Table 2); this estimate is lower than the likely reimbursement and much lower than the typical charge. If the costs of PET were the same as the 2001 Medicare reimbursement for other covered indications for PET (eg, the reimbursement for imaging of the brain for presurgical planning for patients with epilepsy [Health Care Financing Administration Common Procedure Coding System code G0229] would be roughly $2,200, deflated to 1999 dollars [7,8]), the results just described would be more unfavorable for PET.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Sensitivity Analyses: Relationships between Functional Imaging Examinations

 
The hypothetical perfect examination (with a cost of $1,444, equal to that of dynamic susceptibility-weighted contrast-enhanced MR imaging in the base case) dominated all other functional imaging strategies but yielded only 0.0046 more QALY than the standard examination, resulting in an incremental cost-effectiveness ratio of $221,100 per QALY compared with that of the standard examination. The threshold value below which such a perfect examination would yield cost savings was determined by using a simple calculation. All patients incur the cost of the perfect examination, and because the perfect sensitivity and specificity indicate the true disease state for all patients, there is no change in treatment, life expectancy, or costs of care. Consequently, changes in examination cost merely lead to increases or decreases in the mean cost of the perfect examination strategy ($57,876) in a dollar-for-dollar fashion, and QALY gains are unchanged. A decrease in the cost of the perfect examination from our proposed value of $1,444 to $427 would decrease the mean cost of the perfect examination to $56,859 ($57,876 - [$1,444 - $427]), which is identical to the mean cost of the standard examination. Thus, if a perfect examination could be developed for a cost lower than $427, it would yield cost savings. For comparison, the cost of contrast-enhanced CT of the brain (CPT code 70460) was $322 in 2000.

The results of the sensitivity analyses performed to examine relationships between the functional imaging examinations are summarized in Table 2. In scenarios (including the base case) involving the use of PET only in patients with positive standard examination results, the addition of PET was dominated by either the standard examination or dynamic susceptibility-weighted contrast-enhanced MR imaging. In effect, this protocol requires positive results of both PET and the standard examination; thus, any false-negative PET results led to patients with AD being denied both treatment with donepezil and the associated quality-of-life gains. In contrast, when PET was performed in patients with negative standard examination results (essentially, a treatment rule in which all patients with positive results of either examination are treated), the false-negative cases were reduced and the overall effectiveness of the strategy increased (with the assumption that there were no side effects from treating the patients with non–AD-related dementia).

In all scenarios modeled, computed SPECT was dominated by either the standard examination or dynamic susceptibility-weighted contrast-enhanced MR imaging. Relative to the base-case result, the incremental cost-effectiveness ratio of dynamic susceptibility-weighted contrast-enhanced MR imaging decreased with improved drug effectiveness (eg, with drug X, the ratio was $81,700/QALY compared with that of the standard examination) or with longer durations of effectiveness. The inclusion of adverse effects in the patients who were inappropriately treated with donepezil hydrochloride increased the QALY gains generated by all of the functional imaging examinations. However, dynamic susceptibility-weighted contrast-enhanced MR imaging dominated both PET and computed SPECT, regardless of the treatment offered. Estimates of relative cost-effectiveness were similar in the base case and in the scenarios that involved the use of base-case estimates for all inputs, with the exception that (a) HUI2 quality-of-life weights were substituted or (b) a more lenient threshold for the clinical examination was used. Neither varying the cost estimates of imaging examinations by ±50% of the base-case values nor excluding patient and caregiver time costs changed the conclusions in our analysis (17), so the results of these variations are not reported herein.

With the assumption that patients who have non–AD-related dementia experience a decrease of 0.05 QALY per year owing to inappropriate donepezil hydrochloride treatment, the strategy of treating all patients with dementia yielded fewer QALYs (0.6826) at a higher cost ($57,300) than the standard examination; in other words, this strategy was dominated.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
APPENDIX
 REFERENCES
 
The results of this analysis suggest that a combined structural and functional examination, such as dynamic susceptibility-weighted contrast-enhanced MR imaging, may be preferable to PET for the diagnosis of AD. However, the cost-effectiveness ratios of dynamic susceptibility-weighted contrast-enhanced MR imaging have been more than $100,000 per QALY in most analyses: With improvements in therapies or with negative consequences of inappropriate treatment, the incremental cost-effectiveness ratio of dynamic susceptibility-weighted contrast-enhanced MR imaging becomes more favorable (17). Improved nonpharmacologic strategies for AD management could also make functional imaging more useful.

Although the described analysis results might imply that treating all patients who have dementia with cholinesterase inhibitors may be the most reasonable strategy, our aim in this study was not to recommend new treatment guidelines. Furthermore, it is important to note that if adverse effects from inappropriate treatment accrue in patients with non–AD-related dementia, then the strategy of treating all patients with dementia will be dominated by the standard examination and the relative value of functional imaging examinations will increase. Matchar et al (32) have also observed these findings. However, there is a trade-off between side effects and improved effectiveness: As a drug’s effectiveness improves, the relative effect of the adverse effects on the incremental cost-effectiveness ratio begins to wane and the incremental cost-effectiveness ratio of the functional imaging examination increases again. If the strategy of treating all patients with dementia were adopted, then the small proportion of patients who have reversible (ie, curable with treatment other than donepezil hydrochloride) causes of non–AD-related dementia (33–35) would not receive a diagnosis.

We made the simplifying assumption that non–AD-related dementia is reversible and patients with this condition have a quality of life equivalent to that of age-matched control subjects with normal brain function. However, our results are in agreement with those of Matchar et al (32), who modeled patients with non–AD-related dementia as having dementia inexorably progressing to more severe stages. These results indicate that this simplified assumption was not critical to the results.

In our previous study (17), we concluded that dynamic susceptibility-weighted contrast-enhanced MR imaging cost $479,500 (in 1998 dollars) per added QALY, exceeding the value generally considered an acceptable societal willingness to pay for improvements in health. With sensitivity analysis, we demonstrated that future improvements in the effectiveness of therapies to slow, halt, or reverse the cognitive decline in patients with AD will have dramatic effects on the cost-effectiveness of imaging examinations, whereas improvements in imaging examinations alone will not. For example, a lengthened duration of effectiveness of a cholinesterase inhibitor to 4 years (vs the base-case estimate of 18 months) resulted in an incremental cost-effectiveness ratio for dynamic susceptibility-weighted contrast-enhanced MR imaging of $58,930 per QALY. In contrast, with base-case assumptions, a hypothetical perfect examination had an incremental cost-effectiveness ratio of $256,800 per QALY (17).

For patients with inconclusive standard examination results (rather than positive or negative results, as modeled herein), the value of an accurate neuroimaging examination would be highest. In addition to enabling the initiation of therapy, an earlier diagnosis of AD would eliminate uncertainty and allow time to plan for caretaking, legal, and financial arrangements for the patient. Qualitative factors such as these are difficult to include in QALY measures, but they may be important considerations (32).

We did not model a strategy in which all patients underwent both a standard examination and PET, because performing both examinations would yield no additional benefits (but would yield higher costs) other than those from the alternative treatment rule (ie, treat if either standard examination or PET had positive results), which had an incremental cost-effectiveness ratio of $334,200 per QALY (Table 2). In addition to yielding a high cost-effectiveness ratio, this alternative treatment rule resulted in a large number of patients with non–AD-related dementia (54%) being inappropriately treated. The strategy of treating all patients with dementia offered more QALYs at a lower cost than either PET strategy.

The described model assumes that results from the standard examination and PET are independent, which is unlikely to be the case in actual clinical practice. Thus, the probability of PET enabling the diagnosis of a case of AD that was missed at the standard examination is probably overestimated, and the results are slightly biased in favor of PET. The limited data available to stratify sensitivity and specificity by disease stage (due to the progressive nature of AD) or by difficulty of clinical diagnosis (eg, positive, inconclusive, or negative) make the estimation of necessary parameters challenging. Inclusion of the positive correlation between examination results would probably result in PET being less effective than was estimated in our analysis.

The described analyses do not address the primary care or community-based screening examinations in asymptomatic individuals (with or without family medical histories) with AD. In the past 3 years, however, there has been evidence that neuroimaging could enable the identification of patients who are at high risk of developing this disease (36–38).


   APPENDIX
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
REFERENCES
 
Base-case estimates and sensitivity analysis values are listed in Table A1.


View this table:
[in this window]
[in a new window]

  		TABLE A1. Base-Case Estimates and Sensitivity Analysis Values

 


   FOOTNOTES
 
The information cited herein does not necessarily represent the position of the government, and no official endorsement should be inferred.

Abbreviations: AD = Alzheimer disease, FDG = fluorodeoxyglucose, HUI = Health Utilities Index, QALY = quality-adjusted life year

Author contributions: Guarantors of integrity of entire study, P.M.M., G.S.G.; study concepts, P.J.N., P.M.M., G.S.G.; study design, P.M.M., S.S.A., G.S.G.; literature research, P.M.M., S.S.A., E.A.S.; data acquisition, E.A.S., S.S.A., P.M.M.; data analysis/interpretation, P.M.M., S.S.A., G.S.G.; manuscript preparation, all authors; manuscript definition of intellectual content, P.J.N., P.M.M., G.S.G.; manuscript editing, revision/review, and final version approval, all authors.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 REFERENCES
 

  1. Silverman D, Small G, Chang C, et al. Positron emission tomography in evaluation of dementia: Regional brain metabolism and long-term outcome. JAMA 2001; 286:2120-2127.[Abstract/Free Full Text]
  2. Hoffman J, Welsh-Bohmer K, Hanson M, et al. FDG PET imaging in patients with pathologically verified dementia. J Nucl Med 2000; 41:1920-1928.[Abstract/Free Full Text]
  3. Medicare Coverage Advisory publication CAG-00088N. Available at: www.cms.gov/coverage. Accessed July 15 2002.
  4. Hebert L, Beckett L, Scherr P, Evans D. Annual incidence of Alzheimer disease in the United States projected to the years 2000 through 2050. Alzheimer Disease and Associated Disorders 2001; 15:169-173.[CrossRef][Medline]
  5. Brookmeyer R, Gray S, Kawas C. Projections of Alzheimer’s disease in the United States and the public health impact of delaying disease onset. Am J Public Health 1998; 88:1337-1342.[Abstract/Free Full Text]
  6. U.S. General Accounting Office. Alzheimer’s disease: estimates of prevalence in the United States—report to the Secretary of Health and Human Services. General Accounting Office/Health Education and Human Services Publication no. 98–16. Washington, DC: General Accounting Office/Health Education and Human Services, January 1998.
  7. Health Care Financing Administration. Expanded coverage of positron emission tomography (PET) scans and related claims processing changes. Department of Health and Human Services Transmittal Publication no. AB-01–54. Washington, DC: Department of Health and Human Services, April 10 2001.
  8. Health Care Financing Administration. Addendum A: update of calendar year 2001 hospital outpatient payment rates and coinsurance amounts effective April 1, 2001. Available at: www.hcfa.gov/medlearn /apr001a.xls. Accessed January 23 2001.
  9. Rice D, Fox P, Max W, et al. The economic burden of Alzheimer’s disease care. Health Aff (Millwood) 1993; 164-176.
  10. Ernst R, Hay J. The US economic and social costs of Alzheimer’s Disease revisited. Am J Public Health 1994; 84:1261-1264.[Abstract/Free Full Text]
  11. Leon J, Cheng CK, Neumann PJ. Alzheimer’s disease and care: costs and potential savings. Health Aff (Millwood) 1998; 17:206-216.[Abstract]
  12. O’Brien B, Goeree R, Hux M, et al. Economic evaluation of donepezil for the treatment of Alzheimer’s disease in Canada. J Am Geriatr Soc 1999; 47:570-578.[Medline]
  13. Neumann P, Hermann R, Kuntz K, et al. Cost-effectiveness of donepezil in the treatment of mild or moderate Alzheimer’s disease. Neurology 1999; 52:1138-1145.[Abstract/Free Full Text]
  14. Brummel-Smith K. Positron emission tomography for the early diagnosis of dementia (editorial). WJM 1999; 171:294-295.
  15. Scheltens P, Korf E. Contribution of neuroimaging in the diagnosis of Alzheimer’s disease and other dementias. Curr Opinion Neurol 2000; 13:391-396.[CrossRef][Medline]
  16. Small G. Positron emission tomography scanning for the early diagnosis of dementia (opinion). WJM 1999; 171:293-294.
  17. McMahon P, Araki S, Neumann P, Harris G, Gazelle G. Cost-effectiveness of functional imaging tests in the diagnosis of Alzheimer disease. Radiology 2000; 217:58-68.[Abstract/Free Full Text]
  18. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health and medicine New York, NY: Oxford University Press, 1996.
  19. Growdon JH. Advances in the diagnosis of Alzheimer’s disease. In: Iqbal K, Mortimer JA, Winblad B, Wisniewski HM, eds. Research advances in Alzheimer’s disease and related disorders. Chapter 1995; 16:139-151.
  20. Knopman D, DeKosky S, Cummings J, et al. Practice parameter: diagnosis of dementia (an evidence-based review)—report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 2001; 56:1143-1153.[Abstract/Free Full Text]
  21. Morris J, McKeel D, Storandt M, et al. Very mild Alzheimer’s disease: informant-based clinical, psychometric, and pathologic distinction from normal aging. Neurol 1991; 41:469-478.[Abstract/Free Full Text]
  22. Wade J, Mirsen T, Hachinski V, et al. The clinical diagnosis of Alzheimer’s disease. Arch Neurol 1987; 44:24-29.[Abstract]
  23. Mölsä P, Paljärvi L, Rinne J, Rinne U, Säkö E. Validity of clinical diagnosis in dementia: a prospective clinicopathological study. J Neurol Neurosurg Psychiatry 1985; 48:1085-1090.[Abstract]
  24. Silverman D, Phelps M. Evaluating dementia using PET: how do we put into clinical perspective what we know to date? (invited commentary). J Nucl Med 2000; 41:1929-1932.[Free Full Text]
  25. Bosch J, Lester J, McMahon P, et al. Hospital costs for elective endovascular and surgical repairs of infrarenal abdominal aortic aneurysms. Radiology 2001; 220:492-497.[Abstract/Free Full Text]
  26. Feeny D, Furlong W, Boyle M, Torrance GW. Multi-attribute health status classification systems: Health Utilities Index. PharmacoEconomics 1995; 7:490-502.[Medline]
  27. Furlong W, Feeny D, Torrance G, et al. Multiplicative multi-attribute utility function for the Health Utilities Index Mark 3 (HUI3) system: a technical report McMaster University Centre for Health Economics and Policy Analysis Working Paper no. 98–11. Hamilton, Ontario, Canada: McMaster University, 1998.
  28. Neumann P, Sandberg E, Araki S, et al. A comparison of HUI2 and HUI3 utility scores in Alzheimer’s disease. Med Decis Making 2000; 20:413-422.[Abstract/Free Full Text]
  29. Mittmann N, Trakas K, Risebrough N, Liu B. Utility scores for chronic conditions in a community-dwelling population. Pharmacoeconomics 1999; 15:369-376.[CrossRef][Medline]
  30. Grutzendler J, Morris J. Cholinesterase inhibitors for Alzheimer’s disease (review). Drugs 2001; 61:41-52.[CrossRef][Medline]
  31. Foster R, Plosker G. Donepezil: pharmacoeconomic implications of therapy. Pharmacoeconomics 1999; 16:99-114.[CrossRef][Medline]
  32. Matchar D, Kulasingam S, McCrory D, et al. Use of positron emission tomography and other neuroimaging techniques in the diagnosis and management of Alzheimer’s disease and dementia: Duke Evidence-based Practice Center. Available at: www.cms.gov/coverage/download/8b3-ww3.pdf. Accessed September 23 2002.
  33. Duncan BA, Siegal AP. Early diagnosis and management of Alzheimer’s disease. J Clin Psychiatry 1998; 59:15-21.
  34. Freter S, Bergman H, Gold S, Chertkow H, Clarfield A. Prevalence of potentially reversible dementias and actual reversibility in a memory clinic cohort. Can Med Assoc J 1998; 159:657-662.[Abstract]
  35. Miller N, Belkin B, Gold M. Alcohol and drug dependence among the elderly: epidemiology, diagnosis, and treatment. Comprehensive Psychiatry 1991; 32:153-165.[CrossRef][Medline]
  36. Killiany R, Gomez-Isla T, Moss M, et al. Use of structural magnetic resonance imaging to predict who will get Alzheimer’s disease. Ann Neurol 2000; 47:430-439.[CrossRef][Medline]
  37. Bookheimer S, Strojwas M, Cohen M, et al. Patterns of brain activation in people at risk for Alzheimer’s disease. New Engl J Med 2000; 343:450-456.[Abstract/Free Full Text]
  38. Petersen R, Doody R, Kurz A, et al. Current concepts in mild cognitive impairment. Arch Neurol 2001; 58:1985-1992.[Abstract/Free Full Text]
  39. Harris GJ, Lewis RF, Satlin A, et al. Dynamic susceptibility contrast MR imaging of regional cerebral blood volume in Alzheimer disease: a promising alternative to nuclear medicine. AJNR 1998; 19:1727-1732.[Abstract]
  40. Fryback DG, Dasbach EJ, Klein R, et al. The Beaver Dam health outcomes study: initial catalog of health-state quality factors. Med Decis Making 1993; 13:89-102.
  41. Red book Montvale, NJ: Medical Economics Company, 1999.
  42. Bureau of Labor Statistics. Consumer expenditure survey, 1999. Available at: www.bls.gov/csx/1999/standard/age.pdf. Accessed August 2001.
  43. National Vital Statistics Report, vol 47, no. 28. Available at: www.cdc.gov/nchs/products/pubs/pubd/nvsr/47/47-29.htm. Accessed April 18 2003.



This article has been cited by other articles:


Home page
 Br. J. Radiol.Home page
S L PIMLOTT and K P EBMEIER
SPECT imaging in dementia
Br. J. Radiol., December 1, 2007; 80(Special_Issue_2): S153 - S159.
[Abstract] [Full Text] [PDF]

Home page
 BrainHome page
N. L. Foster, J. L. Heidebrink, C. M. Clark, W. J. Jagust, S. E. Arnold, N. R. Barbas, C. S. DeCarli, R. Scott Turner, R. A. Koeppe, R. Higdon, et al.
FDG-PET improves accuracy in distinguishing frontotemporal dementia and Alzheimer's disease
Brain, October 1, 2007; 130(10): 2616 - 2635.
[Abstract] [Full Text] [PDF]

Home page
 NEJMHome page
P. J. Neumann, A. B. Rosen, and M. C. Weinstein
Medicare and Cost-Effectiveness Analysis
N. Engl. J. Med., October 6, 2005; 353(14): 1516 - 1522.
[Full Text] [PDF]

Home page
 RadiologyHome page
G. S. Gazelle, P. M. McMahon, U. Siebert, and M. T. Beinfeld
Cost-effectiveness Analysis in the Assessment of Diagnostic Imaging Technologies
Radiology, May 1, 2005; 235(2): 361 - 370.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2282020915v1
228/2/515    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McMahon, P. M.
Right arrow Articles by Gazelle, G. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McMahon, P. M.
Right arrow Articles by Gazelle, G. S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE