10 Pharmaco-economic issues for diabetes therapy Julia M. Bottomley * PhD Managing Director Amygdala Ltd, The Warren, Willian Road, Letchworth Garden City, Hertfordshire SG6 2AA, UK Frank D. Raymond MBBS, PhD, MD GP Principal Birchwood Surgery, 232-240 Nevells Road, Letchworth Garden City, Hertfordshire SG6 4UB, UK A systematic review was undertaken to analyse pharmaco-economic issues in diabetes, with evi- dence selected on the basis of relevance and immediacy. Pharmaco-economics in diabetes primarily relates to making choices about antidiabetic pharmaceuticals, and this is being influenced by global trends. Trends include increasing numbers of patients with diabetes, with increasing costs of caring for people with diabetes, and an ever-present focus on the costs of pharmaceuticals which are pre- dicted to increase as the pace of development of new medications parallels the increasing incidence of the condition. These developments have influenced the demand for health care in diabetes in the last decade, and will continue to determine this in the coming decade. Recent national experiences are cited to illustrate current issues and to focus specifically upon the challenges facing a raft of new diabetes treatment options now hitting the marketplace, although supported by fewer completed long-term trials. It can be anticipated that these newer agents will be appraised for their cost-effec- tiveness or value for money. Economic analyses for some of the new technologies are summarized; in general, the peer-reviewed publications using well-accepted and validated models have reported that these technologies are cost-effective. Endorsement of any technology in a national setting is not awarded simply because the incremental cost-effectiveness ratio (ICER) falls below the threshold regarded as value for money. In most national observations the reviewers expressed concerns about assumptions used in economic modelling which resulted in the ICERs being deemed optimistic at best, generally highly uncertain, and resulting in the cost-effectiveness appearing better than it really would be in clinical practice. This has often led to the authorities concluding that the price advantage of new technologies over comparators could not be justified, essentially leading to restrictions in use compared to their licence. In general, a paucity of robust evidence on longer-term outcome data together with a lack of health-related quality of life (HRQOL) data collected in a reliable manner * Corresponding author. Tel.: þ44 1462 484611; Fax: þ44 1462 483534. E-mail address: [email protected](J.M. Bottomley) 1521-690X/$ - see front matter ª 2007 Elsevier Ltd. All rights reserved. Best Practice & Research Clinical Endocrinology & Metabolism Vol. 21, No. 4, pp. 657–685, 2007 doi:10.1016/j.beem.2007.08.002 available online at http://www.sciencedirect.com
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Best Practice & Research Clinical Endocrinology & MetabolismVol. 21, No. 4, pp. 657–685, 2007
doi:10.1016/j.beem.2007.08.002available online at http://www.sciencedirect.com
10
Pharmaco-economic issues for diabetes
therapy
Julia M. Bottomley* PhD
Managing Director
Amygdala Ltd, The Warren, Willian Road, Letchworth Garden City, Hertfordshire SG6 2AA, UK
Frank D. Raymond MBBS, PhD, MD
GP Principal
Birchwood Surgery, 232-240 Nevells Road, Letchworth Garden City, Hertfordshire SG6 4UB, UK
A systematic review was undertaken to analyse pharmaco-economic issues in diabetes, with evi-dence selected on the basis of relevance and immediacy. Pharmaco-economics in diabetes primarilyrelates to making choices about antidiabetic pharmaceuticals, and this is being influenced by globaltrends. Trends include increasing numbers of patients with diabetes, with increasing costs of caringfor people with diabetes, and an ever-present focus on the costs of pharmaceuticals which are pre-dicted to increase as the pace of development of new medications parallels the increasing incidenceof the condition. These developments have influenced the demand for health care in diabetes in thelast decade, and will continue to determine this in the coming decade. Recent national experiencesare cited to illustrate current issues and to focus specifically upon the challenges facing a raft of newdiabetes treatment options now hitting the marketplace, although supported by fewer completedlong-term trials. It can be anticipated that these newer agents will be appraised for their cost-effec-tiveness or value for money. Economic analyses for some of the new technologies are summarized;in general, the peer-reviewed publications using well-accepted and validated models have reportedthat these technologies are cost-effective. Endorsement of any technology in a national setting is notawarded simply because the incremental cost-effectiveness ratio (ICER) falls below the thresholdregarded as value formoney. In most national observations the reviewers expressed concerns aboutassumptions used in economic modelling which resulted in the ICERs being deemed optimistic atbest, generally highly uncertain, and resulting in the cost-effectiveness appearing better than it reallywould be in clinical practice. This has often led to the authorities concluding that the price advantageof new technologies over comparators could not be justified, essentially leading to restrictions inuse compared to their licence. In general, a paucity of robust evidence on longer-term outcomedata together with a lackof health-related qualityof life (HRQOL) data collected in a reliable manner
in appropriatepatients andamenable toutility (and hencequality adjusted life yearorQALY)estimationhave resulted in problems for these new drugs at the so-called fourth (cost-effectiveness) hurdle. In thelight of these findings, the implications for generating credible fit-for-purpose cost-effectivenessanalyses of new technologies in diabetes are discussed. Throughout this chapter, the interestedreader is referred to a number of excellent review articles for further details.
Diabetes is a modern-day epidemic and is rightly recognized as a global public healthissue. Trends in the last 10 years are influencing the supply and demand for health carein diabetes (Figure 1) and are focusing the attention of patients, physicians, politiciansand payers of health care alike. All their different perspectives are relevant to phar-maco-economic issues in diabetes.
Health economics applies economic principles to health-care issues which help de-cision-takers make difficult choices when resources are scarce.1 Pharmaco-economicsdetermine the costs and outcomes associated primarily with pharmaceuticals. Theseprinciples are illustrated in Figure 2.
There are several types of pharmaco-economic evaluations available to address dif-ferent decision problems, including cost–utility analyses (CUA), cost-effectivenessanalyses (CEA), and cost–benefit analyses (CBA).1 Figure 2c explains how the burdenof disease is typically described in pharmaco-economic terms: (1) direct costs, i.e.,those related to expenditure for health-care services by the payer or incurred bythe patient themselves to help live with their condition; (2) indirect costs, i.e., thoseoccurring as a consequence of the condition under study, such as missed time fromwork (absenteeism), reduced productivity or on-the-job effectiveness in the workplace(presenteeism) or decreased earning capability or long-term disability with relatedchange in type of unemployment; and (3) intangible costs, (i.e., the impact of the dis-order on the health-related quality of life (HRQOL) of the patient).1
Making decisions in health care rely heavily on evidence from randomized con-trolled trials (RCTs), but these invariably leave information gaps. Decision-makershave therefore increasingly used computer modelling techniques to structure the
Figure 1. Global trends and their impact on pharmaco-economics of diabetes.
Pharmaco-economic issues for diabetes therapy 659
evidence on clinical and economic outcomes to facilitate decision-making.2 Modelssimulate reality, but in practice are invariably simplifications thereof. Features of diabe-tes also pose unique challenges for economic modelling3, but the current diabetesmodels are more complex, validated, fit for purpose, and increasingly accepted by de-cision-makers. The conclusion of each economic evaluation is conditional upon theunique data and assumptions underpinning it. Pharmaco-economics is now morethan ever before at the leading edge of thinking in terms of securing the most rationaland efficient use of scarce health-care resources for diabetes.
INCREASING NUMBERS OF PEOPLE WITH DIABETESAND ‘PRE-DIABETES’
The last decade has seen a meteoric rise in the incidence and prevalence of diabetesworldwide.4 In the USA, from 1995 through 2005 the numbers diagnosed with diabe-tes increased from 8 to 15.8 million5; including undiagnosed cases, this is estimated toexceed 20 million people or 7% of the population6, and this trend is set to continue.Type-2 diabetes accounts for over 90% of cases of diagnosed diabetes, and is associ-ated with older age, obesity, physical inactivity and race/ethnicity.7 The alarming in-creases are due to type-2 diabetes, and its worldwide prevalence is predicted to
Figure 2. A summary of health economics; (a) principles; (b) health-care value equation; (c) costs. HRQOL,
health-related quality of life.
660 J. M. Bottomley and F. D. Raymond
reach 300 million in 2025, with a further 418 million people with impaired glucose tol-erance (IGT).4,8 An added focus at national government level is that the obesity pan-demic occurring in major developed and developing markets has contributed to theincrease in type-2 diabetes in adults7,9 and children.10,11 As a consequence, type-2 di-abetes is the focus of this chapter. However, for completeness, the authors also discussfindings of relevance in type-1 diabetes and pre-diabetic conditions (‘pre-diabetes’).
INCREASING SOCIETAL BURDEN OF DIABETESAND ‘PRE-DIABETES’
The socioeconomic consequences of diabetes and ‘pre-diabetes’ are huge.9,12 Peoplewith type-2 diabetes can anticipate a life expectancy shortened by 10 years comparedwith the general population, due mainly to an increased risk of cardiovascular (CV)death, myocardial infarction (MI) or stroke.13,14 Type-2 diabetes has been defined asa state at risk of premature CV death associated with hyperglycaemia and microvas-cular disease15, because more than 65% of deaths are from CV causes.16 The associ-ated significant comorbidity and mortality is caused by the development of long-termmicrovascular and macrovascular complications. HRQOL is an important issue forpeople with type-2 diabetes; it decreases with disease progression and diabetic com-plications.17–20 The progression of diabetes demands high levels of medical care, andthe management of patients is inevitably costly. It is well established that the hospitalcosts of the complications is the overriding cost factor21–27, with the majority of hos-pitalizations due to CV complications.28 The Cost of Diabetes in Europe Type-2 Study(CODE-2) reported that 72% of the patients had at least one complication, of whicha third had macrovascular disease.29 The cost of care was up to 250% higher in peoplewith complications compared to those without complications. The findings of the fullcosts of care of people with type-2 diabetes in the UK were in accord with those ofthe overall CODE-2 study and those of the cross-sectional T2ARDIS survey (Type-2Diabetes Accounting for a major Resource Demand In Society).22,30 In a study focusingon drug costs, patients with type-2 diabetes accounted for just over 7% of the totaldrugs budget31, with excess prescribing costs reflecting higher use of CV drugs thanin the general population.
Diabetes also has significant economic effect on the person and their family.32 Lostproduction caused by sickness absence from work, early retirement, and prematuremortality carry high costs to the public purse. Currently almost 50% of Americanadults with diabetes are aged between 45 and 64 years5, while over 60% of the newcases of diagnosed diabetes in 2005 were aged 20–59 years6 and thus in the most eco-nomically productive age group. Recent research has demonstrated the hugely delete-rious impact of CV risk factor clusters on social productivity in the USA.33
Furthermore, 80% of the world’s diabetes resides in lower-income nations, posingan even greater burden on developing economies.34 Total annual costs of diabetesin the USA in 2002 were estimated at $92 billion for direct costs (costs of medicalcare and services) and $40 billion for indirect costs (costs of short-term and perma-nent disability, work loss, and premature death).6,35 People with diabetes had medicalexpenditures some 2.4-fold higher than expenditure incurred by matched populationwithout diabetes.36 Similar findings were reported in three other studies. In the UK intype-1 diabetes36 and Sweden (all diabetes)37, indirect costs comprised more than halfof the total costs. Holmes et al32 noted that the loss of earnings experienced by type-2diabetes patients and their carers was significant, and a strong association was found
Pharmaco-economic issues for diabetes therapy 661
between patients’ loss of earnings and the presence of complications. Care providedby informal carers is an important aspect of total care in diabetes and other chronicconditions, and the importance of this is now increasingly being acknowledged.38
There are thus economic reasons together with human factors for seeking to reduceor delay the complications of diabetes.
IMPACT OF TREATMENT OPTIONS ON COMPLICATIONS AND COST
The aim of treating type-2 diabetes is to delay or prevent the complications, and it iswell known that improving glycaemic control and associated CV risk factors substan-tially reduces the risk39–50 and associated costs51–65 of long-term complications. Im-portantly, treatment of risk factors has been associated with a positive impact onHRQOL in newly diagnosed patients19,66 as well as people with more advancedtype-2 diabetes.18,20
Effective glycaemic control is also associated with decreased incidence67,68 andcosts69,70 of complications in patients with type-1 diabetes. Several studies have shownthat delay or possible prevention of progression of ‘pre-diabetes’ to diabetes can beachieved with lifestyle changes and oral antidiabetic agents71–77 and that sustained gly-caemic control may slow the progression of type-2 diabetes.51,78 Analyses of costs andcost-effectiveness of prevention of type-2 diabetes have found lifestyle modificationand medication to be cost-effective.79–83
PHARMACO-ECONOMICS AND TYPE-2 DIABETES
The growing interest in the optimal allocation of health-care funding in prevention oftype-2 diabetes and its complications is reflected in four reviews of economic evalua-tions in diabetes. The first three had different perspectives and inclusion criteria andspanned different periods.27,84,85 The latest systematic review86 focused on studiesconcerning the more costly macrovascular complications and drew a number ofconclusions in terms of efficient use of health-are expenditure (see Figure 3). Strict
Figure 3. Cost-effective use of resources in type-2 diabetes. Adapted from Vijgen et al.86
662 J. M. Bottomley and F. D. Raymond
blood-pressure control is more cost-effective than less-strict control. Primary and sec-ondary (i.e. screening or early detection and subsequent treatment) prevention ishighly cost-effective (but results were based on very few studies). Medications toreduce weight and HbA1c simultaneously were cost-effective versus conventional in-terventions (but this was based on only three trials). However, the individual resultsfor medications to reduce weight, hyperglycaemia and hypercholesterolaemia showedwide variation, with the newer (more expensive) agents in general producing high incre-mental cost-effectiveness ratios (ICERs), indicating that they were not value for money.
Vijgen et al86 suggested that the newer technologies deserve attention from policymakers, and that further economic analyses are needed, especially studies which com-pare all relevant medications with each other to establish which drug, or combinationof drugs, is most cost-effective. International comparison of economic evaluations wasacknowledged to be difficult, and the authors called for more attention to be devotedto the assessment of uncertainty in cost-effectiveness studies based on modelling.
A recent report agreed that screening was cost-effective, noting that the very lowcost of statins was an important factor in these conclusions.87 Icks and colleagues an-alysed the clinical and cost-effectiveness of primary prevention of type-2 diabetes ina ‘real-world’ setting.88 They concluded that implementing prevention programmesin clinical practice is costly and may not prevent high numbers of diabetes, that accom-panying costs such as staff education and especially intervention participation and ad-herence should be considered, and that further studies are warranted to determinethe optimal prevention programme in clinical practice.
INCREASED CHOICE: NEW TREATMENT OPTIONS IN DIABETES
Metformin and sulphonylurea as monotherapy or combination therapy have tended tobe the mainstay of first- and second-line therapy in type-2 diabetes, although some cli-nicians are increasingly advocating use of at least two drugs to obviate the monother-apy failure that accompanies long-term metformin in most patients. Stringentglycaemic control is important to the management of type-2 diabetes; reducingHbA1c levels down to the normal range is considered the Holy Grail of diabetestreatment.7,89
Despite a plethora of national guidelines, an increasing evidence base points to poorcontrol despite intensification of therapy. Many people with type-2 diabetes who areprescribed multiple oral agents and/or insulin in routine practice are still receiving in-adequate monitoring and achieving poor glycaemic control; these failures are seenacross races and ethnic groups.90–96 In general, oral antidiabetic drugs used in combi-nation are able to achieve reasonable HbA1c levels for around 10 years.97 Thereafter,the progressive nature of type-2 diabetes means that patients may need to move toinsulin therapy, which may present a management challenge for both patients and phy-sicians. The key message remains, that intensive management is still required to reducethe risk of complications.
The pace of development of new medications to treat diabetes has paralleled theincreasing incidence of the condition. In recent years a myriad of new technologieshas emerged from clinical development, offering the potential to address someof the underlying pathophysiological defects which currently available agents do notcorrect. These new modalities potentially offer valuable solutions for patients andsociety98–100, but at additional cost.
Pharmaco-economic issues for diabetes therapy 663
There are now nine classes of medication available to treat type-1 and type-2 dia-betes. Table 1 summarizes the distinctive features of the classes now available; thenewer drugs can expect to undergo appraisal for their value for money. These includeinhaled insulin, the dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like-peptide-1(GLP-1) analogues and amylin analogues. (Table 1 excludes fixed-dose combinationsof different classes of drug but describes a number of other candidates in companypipelines.) The expansion from use of insulin, metformin, sulphonylurea and a-glycosidase inhibitors to classes including postprandial regulators, incretins, thiazolidi-nediones, inhaled insulins, amylin analogues and DPP-4 inhibitors has not beensupported by completed longer-term studies for all these classes.98–100
At face value, we could conclude from Table 1 that there are now more options forpatients and physicians, which is good. It cannot confirm which drug is better than an-other. The challenge for clinicians, payers, and other decision-makers is how to chooseamong these different drug classes.
COST-EFFECTIVENESS OF NEWER TECHNOLOGIES FOR DIABETES
This chapter concentrates on the newer (more expensive) classes of drugs with eco-nomic analyses in type-1 and type-2 diabetes, summarized in Tables 2 and 3, respectively.We have identified recent evidence relating to long-acting insulin analogues, inhaledinsulins (Exubera�), incretins (exenatide, Byetta�) and rimonabant (Acomplia�).The appearance of these newer medicines is not yet supported with longer-termclinical trials, and the published economic literature to support these technologies islimited.
These recent economic analyses were reviewed to compare key aspects of thestudies. Table 2 (type-1 diabetes) and Table 3 (type-2 diabetes) summarize countryof origin, year of study, perspective, comparators, the economic model used to fore-cast long-term costs and outcomes, the quality adjusted life years (QALYs) gained, andthe ICERs reported. To facilitate comparison across these studies the authors con-verted the reported ICERs into cost-effectiveness ranges (V per QALY gained, year2007 values), where a commonly quoted ‘good-value-for-money’ threshold isV50,000 per QALY gained in Western Europe.
Long-acting basal insulin analogues
Insulin analogues have been reported to be cost-effective in comparison to their re-spective basal or bolus human insulin. A UK modelling analysis in type-1 diabetes pro-jected an ICER of £19,285 per QALY gained for insulin detemir compared with neutralprotamine Hagedorn (NPH) insulin-based basal-bolus therapy.101 A second analysis,also in type-1 diabetes, predicted a cost per QALY of £4825 for insulin aspart versushuman soluble insulin.102 McEwan and colleagues103 recently concluded that insulinglargine treatment is highly cost-effective compared with NPH insulin in people withtype-1 diabetes whether the outcome of interest is reduced occurrence of hypogly-caemia or improvement in HbA1c (Table 2). In a second modelling study, insulin glar-gine was cost-effective for all patients with type-2 diabetes104, and not just those athigh risk of hypoglycaemic events (Table 3). Palmer and colleagues recently reportedthe long-term clinical and economic outcomes of analogue insulin (insulin aspartplus insulin determir) compared to human insulin in a basal-bolus regimen in patientswith type-1 diabetes.105 The underlying clinical trial demonstrated a 0.22% greater
of type-1 and type-2 diabetes.
n Inhaled
insulin
Amylin
analogues
DPP-4
inhibitors
Incretin
mimetics
1.5 0.5e1.0
(with insulin)
0.5e0.9 1e1.4 (in
combination)
ion Inhaled Injection Oral Injection
+++ Gain ++ Loss Neutral Loss
++ + +++
Yes Yes Minimal Increased
risk in
combination
with SUs
No No data No data
lished
eases
ovascular
QOL and
satisfaction
Improved
adherence?
No need
to adjust
dose relative
to HbA1c
levels
Table 1. Impact of drugs as monotherapy for the treatment
Variable Class of druga
SU Meglitinides Biguanides AGI TZD Insuli
HbA1c
decrease (%)
1e1.5 0.5e1.5 1.5 0.5e0.8 0.8e1.0 �2.5
Route Oral Oral Oral Oral Oral Inject
Weight impact Gain ++ Less than SU? Loss Neutral Gain +++ Gain
Gastrointestinal
effects
+ ++ +++ +
Hypoglycaemia Yes Less than SU? No No No Yes
Oedema + +++
Macrovascular
risk reduction
No Yes No data Yes No
Prevent
diabetes
Yes Yes Rosiglitazone
Pioglitazoneb
(troglitazone)c
Other
advantagesWell
established
Decreases
microvascular
risk
Targets
postprandial
glycaemia
Well established
Decreases
microvascular
risk
Non-glycaemic
benefits
Targets
postprandial
glycaemia
Non-
glycaemic
benefits
Glycaemic
durability
Well
estab
Decr
micr
risk
Injection site
reactions
Patient fear
or reluctance
to inject
Worsening
retinopathy in
patients with
retinopathy if
HbA1c
reductions
too rapid
Cough
Use in
smokers and
other
respiratory
conditions
More long-
term data
needed
Bulky
device
Requires
dose
adjustment
to prevent
hypos
No long-
term data
No long-
term data
Very rapid
acting
Short acting
Intermediate
and long-acting
Biphasic
Exubera
Several
others in
development
Symlin1Sitagliptin2
Vildagliptin3
SYR3224
Saxagliptin5
PHX11496
R14387
PSN93018
BI-13569
Exenatide10
Liraglutide11
DAC:GLP-112
Exenatide
LAR13
AVE 000100
(ZP10)14
(continued on next page)
Other
disadvantagesHyperinsulina-
emia
b-cell
exhaustion?
Impaired kid-
ney function
More long-
term data
needed
Complex
dosing (tid)
Hyperinsulina
emia
No long-term
data
Impaired liver
function
Lactic acidosis
(rare)
Contraindicated
with renal
impairment or
heart failure
needing
medication
More long-
term data
needed
Complex
dosing (tid)
Impaired
liver
function
LFT
monitoring
Slow onset
of action
More long-
term data
needed
Risk of
fracture
Impaired
cardio
pulmonary
functionProductsd
Glibenclamide
Gliclazide
Glimepiride
Glipizide
Nateglinide
Repaglinide
Metformin (M) Acabose
miglitol
Pioglitazone
Rosiglitazone
Several in
development
Inhaled
insulin
Amylin
analogues
DPP-4
inhibitors
Incretin
mimetics
e
Type-1 diabetes:
in combination
with long or
intermediate
Insulin
Type-2 diabetes:
Alone or in
combination
with oral medi-
cation or long
acting insulins
Type-1
diabetes:
Adjunct to
mealtime
insulin
Type-2
diabetes:
adjunct to
mealtime
insulin� SU
or M
Monotherapy
Combination
with M
or TZD
Combination
with SU
and/or M
U, sulphonylurea; TZD, thiazolidinedione;
etact) or metformin & rosiglitazone (Avandamet)) not included
Table 1 (continued)
Variable Class of druga
SU Meglitinides Biguanides AGI TZD Insulin
FDA Approval
statusMonotherapy
Combination
with I, M,
TZD, ACI
Monotherapy
Combination
with M
Monotherapy
Combination
with I, SU, TZD,
Non-SU
secretagogues
Monotherapy
Combination
with SUs
Monotherapy
Combination
with I
(piogli-
tazone)e,
SU, M
Monotherapy
Combination
with SU,
M TZD
(pioglitazone)
ACI, a-glucosidase inhibitors; DPP-4, dipeptidly pepdidase-4; LFT, liver function test; M, metformin; QOL, quality of life; S1 Symlin (pramlitide acetate) (Amylin), FDA approved;2 Sitagliptin (Merck), approved;3 Vildagliptin (Novartis), European regulatory decision anticipated end 2007; in discussions with FDA;4 SYR322 (Alogliptin) (Takeda): Phase III;5 Saxagliptin (Bristol-Myers Squibb/Astra Zeneca): Phase III;6 PHX1149 (Phenomix): Phase II;7 R1438 (Roche): Phase II;8 PSN9301 (Prosidion): Phase II;9 BI-1356 (Boehringer-Ingelheim): Phase II/III;
a Fixed dose combinations excluded.b Pioglitazone reduced incidence of type-2 diabetes in women with prior gestational diabetes.c Troglitazone withdrawn due to liver toxicities.d Combination products (e.g. metformin & Glibenclamide (Glucovance); metformin & pioglitazone (Comp
for simplicitye An indication for rosiglitazone-insulin combination is under consideration by the EMEA.
Pharmaco-economic issues for diabetes therapy 667
reduction in HbA1c levels with the analogue group (P¼ 0.001).106 Their results pro-jected an ICER of £2500 per QALY gained, and suggest that by using analogue insulinsin combination, this can increase their cost-effectiveness, for example by reducing theICER for insulin detemir compared to human insulin from £19,285101 to £2500.105
Table 2 cites ICERs for these insulin analogues which are considered to be value formoney when benchmarked against a willingness to pay (WTP) threshold of V50,000per QALY gained.
Inhaled insulin (Exubera�)
Two studies have reported that inhaled insulin is a cost-effective alternative in patientswith uncontrolled type-1 and type-2 diabetes. The UK study evaluated the cost-effec-tiveness of Exubera in patients with uncontrolled type-1 diabetes (on a premix regi-men) and type-2 diabetes (uncontrolled on oral antidiabetic drugs) using thevalidated EAGLE model107 to simulate the long-term clinical and cost-effectivenessover a 20-year time horizon (Tables 2 and 3). The authors concluded that at aWTP threshold of �£30,000 per QALY gained108 inhaled insulin was cost-effectivein all patients with type-1 diabetes and in 92% to 95% of patients with type-2 diabe-tes.109 A Swedish study used the same economic model to compare the cost-effective-ness of inhaled insulin with standard treatment in patients with uncontrolled type-1 ortype-2 diabetes.110 At a WTP threshold of SEK 500,000 per QALY gained, inhaled in-sulin appeared to be a cost-effective option in patients with uncontrolled type-1 ortype-2 diabetes (see Tables 2 and 3), although the results were sensitive to the levelof treatment-associated utilities.
Exenatide (Byetta�)
Along with modest reductions (0.5–0.9) in HbA1c, a statistically significant weight loss of2–3 kg from baseline is typically observed in exenatide clinical trials.111,112 However, theneed for twice-daily injections and mild to moderate nausea might be challenging forsome patients (Table 1). Table 3 describes three cost-effectiveness studies of the adjunc-tive use of exenatide twice daily compared with standard add-on therapies in patientswith type-2 diabetes who achieved inadequate blood glucose control despite therapywith metformin and/or a sulphonylurea. Watkins and colleagues demonstrated howthe validated CORE diabetes model69,113 aided US formulary decision-making for exena-tide in the absence of long-term outcomes data; all patients in the simulation receivedmetformin, and the authors modelled head-to-head comparisons of exenatide versusonce-daily insulin glargine, pioglitazone 30 mg, generic glyburide (glibenclamide)15 mg, or no additional treatment.114 An analysis using the same model evaluated thelong-term clinical and economic outcomes of exenatide versus insulin glargine in a UKsetting.115 Additional exenatide therapy was predicted to increase QALYs comparedwith other comparators, and this translated into ICERs which were deemed to becost-effective in the US and the UK. These two modelled analyses using the same modelpredicted lower QALY gain in the UK analysis (Table 3). This could reflect the use of dis-counting or it could be accounted for by other assumptions used in the analysis.
In contrast, an indirect comparison approach to a within-trial economic analysis116
concluded that exenatide did not appear to be cost-effective for the treatment of type-2 diabetes in the USA when compared with an intermediate and a long-acting insulin(glargine). However this short-term analysis (only 24 weeks) described costs per
e-1 diabetes treatments.
ount
(%)
Model to
project
long-term
outcomes
QALYs
gained
versus
comparator
Reported ICER
(cost per
QALY gained)
Converted
ICER range
(V, 2007)a
EAGLE 0.024 �£8510 10,000e20,000
EAGLE Not
reported
SEK 38,948 1000e10,000
DES 0.12e0.14 £7400e8810b 10,000e20,000
DES 0.14e0.34 £3190e9770c 10,000e20,000
CDM 0.66 £2500 1000e10,000
of glycaemic control and long-term effects of diabetes model; HIS,
H, neutral protamine Hagedorn insulin; QALY, quality adjusted life
thors to make the ICERs more comparable (V50,000 per QALY
668
J.M
.B
otto
mley
and
F.D
.R
aymond
Table 2. Cost-effectiveness analyses of new classes of typ
Study New therapy Comparator Country
(year of costing,
perspective)
Time horizon
(years)
Disc
rate
Plun-Favreau et al109 Exubera Injected basal
bolus regimen
UK (2005, NHS) 20 3.5
Neovius and Pruz110 Exubera Injected basal
bolus regimen
Sweden
(2005, Swedish
Healthcare)
20 3
McEwan et al103 Insulin glargine NPH insulin UK (2005, NHS) 40 3.5
Insulin glargine NPH insulin UK (2005, NHS) 40 3.5
human soluble insulin; ICER, incremental cost-effectiveness ratio; NHS, National Health Service; NP
year.a Reported ICERs from the different studies were converted into V per QALY gained by the au
gained is regarded as cost-effective in Western Europe).b Based on reductions in hypoglycaemic events.c Based on improvement in HbA1c.
Pharmaco-economic issues for diabetes therapy 669
surrogate outcome measures which did not facilitate comparison with the above-mentioned analyses which reported cost per QALY gained.
Tables 2 and 3 indicate that the majority of these published ICERs for the new diabetestreatment options fall within the range regarded as value for money in Western Europe.
INCREASING SCRUTINY OF THE COST-EFFECTIVENESSOF DIABETES TREATMENTS
There are increasing demands by decision-makers in different countries, particularly inthe Health Technology Assessment (HTA) arena, to justify the cost-effectiveness ofnew and existing technologies in health care in real-world settings (Figure 1). In orderto make funding or reimbursement decisions, payers or national agencies such as theNational Institute for Health and Clinical Excellence (NICE) require robust and rele-vant outcome data, including HRQOL and costs to address patient experiences withtreatment to make recommendations on the clinical and cost-effectiveness of new andexisting technologies. HRQOL outcomes are a pivotal element in the evidence to sup-port new pharmaceuticals117 to generate the gold standard outcome in economic eval-uation of cost per QALY gained. Importantly, this enables benchmarking of onetechnology versus another in terms of value for money.
Published NICE appraisals are being used as international benchmarks.118 Impor-tantly, in a global market where NICE enjoys a formidable influence its requirementsprovide a useful example of the health economic issues facing new technologies inmany countries. An economic approach which persuades NICE has credibility to influ-ence pricing and reimbursement and HTA in Europe, Australia, and Scandinavia, as wellas the USA and Canada, with appropriate national adaptation.
Given the high costs of diabetes, the prospect of newer, more expensive agents(with more favourable pharmacological therapeutic indices than drugs already inuse), and the consequent increase in spend on pharmaceuticals, it is important to rig-orously compare the relative benefits and costs of these treatments with existingtreatment regimens before newer medicines gain access to a market and before treat-ment policies are changed. Tables 2 and 3 describe a raft of peer-reviewed economicanalyses reporting that all these technologies can be regarded as value for money indiabetes according to WTP thresholds. The next section compares the economic anal-yses with the decisions reached regarding access of the drug to the market. Theauthors therefore posed the questions:
1. How have recent pharmaco-economic studies of newer diabetes drugs influencedthe use of these agents in various countries? Did the decision-makers agree withthe authors conclusions?
2. Were the economic evaluations fit for purpose to satisfy the demands of the vari-ous decision-makers tasked with determining the cost-effectiveness (or value formoney) of these new drugs?
3. What are the implications for pharmaco-economics in diabetes?
Reimbursement recommendations for new treatments in diabetes
Recent HTA recommendations enable an appreciation of the current issues and ex-pectations of pharmaco-economic analyses in diabetes and inform future data needs.
betes treatments.
Model to
project
long-term
outcomes
QALYs
gained
versus
comparator
ICER
(cost per
QALY gained)
Converted
ICER range
(V, 2007)a
EAGLE 0.021 �£24,285 20,000e50,000
EAGLE 0.021 �£24,555 20,000e50,000
EAGLE Not
reported
SEK 151,186 10,000e20,000
EAGLE Not
reported
SEK 178,106 10,000e20,000
EAGLE Not
reported
SEK 162,294 10,000e20,000
EAGLE Not
reported
SEK 265,376 20,000e50,000
EAGLE Not
reported
SEK 232,442 20,000e50,000
EAGLE Not
reported
SEK 183,132 20,000e50,000
DES 0.111b £10,027 10,000e20,000
DES 0.111c £13,921 10,000e20,000
CDM 0.44 £22,420 20,000e50,000
670
J.M
.B
otto
mley
and
F.D
.R
aymond
Table 3. Cost-effectiveness analyses of new classes of type-2 dia
Study New therapy Comparator Country
(year of costing,
perspective)
Time horizon
(years)
Discount
rate (%)
Plun-Favreau et al109 Exubera Injected basal
regimen
UK (2005, NHS) 20 3.5
Exubera Injected premix
regimen
UK (2005, NHS) 20 3.5
Neovius and Pruz110 Exubera Injected basal
bolus regimen
Sweden
(2005, Swedish
Healthcare)
20 3
Exubera Intensifying to
basal insulin
Sweden
(2005, Swedish
Healthcare)
20 3
Exubera Intensifying to
mix-insulin
Sweden
(2005, Swedish
Healthcare)
20 3
Exubera Intensifying to
mix-insulin
Sweden
(2005, Swedish
Healthcare)
20 3
Exubera Intensifying to
basal-bolus
Sweden
(2005, Swedish
Healthcare)
20 3
Exubera Intensifying to
basal bolus
Sweden
(2005, Swedish
Healthcare)
20 3
McEwan et al104 Insulin glargine NPH insulin UK (2005, NHS) 40 3.5
Insulin glargine NPH insulin UK (2005, NHS) 40 3.5
Ray et al115 Exenatide Insulin glargine UK (2004; NHS) 35 3.5
30 Not
specified
CDM 1.3 $US13,000 10,000e20,000
30 Not
specified
CDM 1.0 $US32,000 20,000e50,000
30 Not
specified
CDM 1.7 Dominant Dominant
30 Not
specified
CDM 1.2 $US16,000 10,000e20,000
d) <0.5 Assumed 0 NA Not reported Not reportedd Not reported
d) <0.5 Assumed 0 NA Not reported Not reported* Not reported
conomic assessment of glycaemic control and long-term effects of diabetes model; ICER,
uality adjusted life year; NPH, neutral protamine hagedorn; NA, not applicable.
ALY gained by the authors to make the ICERs more comparable (V50,000 per QALY
tion over the 24 week studies.
Pharm
aco-eco
nom
icissu
esfo
rdiab
etesth
erapy671
Watkins et al114 Exenatide Insulin glargine US
(2005, private payer)
Glyburide US
(2005, private payer)
Pioglitazone US
(2005, private payer)
No treatment US
(2005, private payer)
Edwards et al116 Exenatide Insulin NPH US (2005, not specifie
incremental cost-effectiveness ratio; NHS, National Health Service; QALY, qa Reported ICERs from the different studies were converted into V per Q
gained is regarded as cost-effective in Western Europe).b Hypoglycaemic benefit only.c HbA1c benefit only.d ICERs described as cost per HbA1c reduction or cost per weight reduc
672 J. M. Bottomley and F. D. Raymond
HTA experience or other national cost-containment efforts to institute cost-effectiveprescribing for a number of these newer diabetes agents are limited; however, thereare some important recent national observations, and these are summarized next.
Long-acting basal insulin analogues
In 2002, NICE concluded that insulin glargine was cost-effective in type-1 diabetes butonly marginally cost-effective in people with type-2 diabetes.119,120 Consequently insu-lin glargine was recommended in type-1 diabetes but its use was restricted in type-2diabetes to those patients at highest risk of hypoglycaemia or those who would espe-cially benefit from its once-daily administration. At the time, the cost-effectiveness wasinfluenced most strongly by an assumption about the reduction in HRQOL associatedwith increasing frequency of symptomatic hypoglycaemic events (from an unpublishedsource). NICE considered that the utility gains associated with reducing the fear of hy-poglycaemia and other assumptions used in the manufacturer’s analysis biased the re-sults in favour of insulin glargine. Since then there has emerged a better appreciation ofthe anxiety and fear of patients associated with hypoglycaemic events; this has beenreported to be associated with severe hypoglycaemia as opposed to frequency ofsymptomatic hypoglycaemia.121
Similar recommendations for use of both insulin glargine and insulin detemirwere issued in Scotland122 and more recently in Australia.123 Reviewers expressedconcerns about assumptions used in the economic modelling which resulted in theICERs being regarded as highly uncertain. This led the authorities to conclude thatthe price advantage of long-acting insulins over comparators was not justified, therebyrestricting its use to type-1 diabetes but not in type-2 diabetes.
Inhaled insulin (Exubera�)
Exubera is a rapid-onset, short-acting mealtime insulin offering a well-tolerated and ef-fective alternative and/or addition to injectable insulins. Its unique delivery can reducethe number of times per day that a patient must self-inject and addresses a number oflimitations of current insulin-containing regimens. As such, this could facilitate earlieruptake and intensification of insulin treatment in type-2 diabetes and earlier intensifi-cation in type-1 diabetes, and could be expected to lead to greater patient adherencewith their insulin treatment.
The Scottish Medicines Consortium (SMC) reviewed Exubera and did not recom-mend it for use within NHS Scotland in type-1 or type-2 diabetes because the eco-nomic case was not demonstrated.124 While the SMC acknowledged that thestrengths of the economic analysis were in using a well-validated model for epidemi-ology and adverse clinical events, together with their associated costs and utilities,they had four main concerns:
� generalizability of the clinical comparators to Scottish practice, and particularly as-sumptions on delays in switching to more intensive insulin therapy in the standard-care arm;� assumptions about glycaemic control for patients on inhaled insulin versus some not
receiving insulin;� assumed gains in utility associated with inhaled insulin; and� uncertainty surrounding the dose (and cost) of inhaled insulin.
Pharmaco-economic issues for diabetes therapy 673
NICE subsequently reviewed the clinical and cost-effectiveness of Exubera in diabe-tes.125 Inhaled insulin was not recommended for the routine treatment of people withtype-1 or type-2 diabetes mellitus despite widespread pressure from expert cliniciansand patient groups that this product can revolutionize diabetes treatment. In contrastto Scotland, use was restricted to patients with poor glycaemic control and who eitherhad marked and persistent fear of injections or severe and persistent injection siteproblems. In coming to this recommendation, NICE noted the following126:
� none of the published trials compared Exubera with short-acting injectable ana-logues, which would have been a better comparison;� the only significant difference between inhaled and soluble insulin in the trials was
patient preference;� the manufacturers of Exubera argued that this patient preference could lead to im-
proved control in some type-1 patients through improved compliance, and in sometype-2 patients poorly controlled on oral agents, because a switch to insulin therapywould be more acceptable if people could use inhaled rather than injected insulin;� while the EAGLE model was validated and reputable, the results depend upon the
underlying assumptions fed into the model;� key assumptions were the size of the gain in utility from inhaling rather than injecting
insulin, and the effect of having an inhaled option on the willingness to start insulinamongst patients with poor control on antidiabetic drugs; and� NICE considered that the assumptions used in the industry submission made the
cost-effectiveness appear better than it really would be.
Exenatide (Byetta�)
Byetta was recently appraised by the SMC and accepted for restricted use within NHSScotland for type-2 diabetes in combination with metformin and/or sulphonylureas inpatients not achieving adequate glycaemic control on maximally tolerated doses ofthese oral therapies.127 The SMC noted that exenetide had shown non-inferiorityto two insulin regimens and had a beneficial effect on weight. It is restricted to useas an alternative to insulin in patients in whom insulin would be the next treatmentoption. The submitted cost utility analysis was based on a 52-week RCT which com-pared exenetide to biphasic insulin aspart128 and used the CORE diabetes model toproject outcomes over a 10-year time horizon. Patients in the exenetide arm of themodel were assumed to be treated with exenetide for 3 years and then switchedto biphasic insulin aspart. The analysis assumed that any weight loss by exenetide pa-tients would be regained when the patient switched to insulin, and also that their bodymass index (BMI) would rise to the level of insulin-treated patients during year 1 of theswitch. The cost per QALY for this scenario, using utility values from the CODE-2study18, was £6790.127 The SMC expressed concern that the model did not compareexenetide to cheaper forms of insulin, but acknowledged that sensitivity analyses ver-sus biphasic human insulin suggested that exenetide is likely to remain cost-effective.
Sitagliptin (Januvia�)
The authors did not identify any published economic evaluations for sitagliptin. How-ever, recommendations for its use based on available clinical evidence have been issued
674 J. M. Bottomley and F. D. Raymond
in the USA. ‘Because sitagliptin modestly reduces HbA1c, has no long-term safety and out-comes data, and is significantly more costly compared to the formulary agents’, the VHAPharmacy Benefits Management Strategic Healthcare Group and Medical AdvisoryPanel, in promoting cost-effective prescribing, recommended that stringent criteriabe met (limiting the patient pool, start and stop rules) when sitagliptin is used asmonotherapy or in combination.129 Furthermore, if sitagliptin is prescribed butHbA1c goals are not met after 3–6 months of therapy, it should be discontinued andinsulin initiated. Nevertheless, many clinicians consider an alternative oral agent orexenetide to be more appropriate than insulin.
Rimonabant (Acomplia�)
Acomplia was appraised by the SMC and was not recommended for use as an adjunctto diet and exercise for the treatment of obese or overweight patients with an asso-ciated risk factor such as diabetes or dyslipidaemia.130 The SMC noted that rimona-bant was associated with a mean weight reduction of around 4–5 kg over that ofplacebo; however, the weight was generally regained within 1 year of stopping treat-ment. The economic case was not demonstrated, primarily because it was not clearthat the comparator reflected current practice in Scotland.
In 2007 in Germany, Acomplia was not deemed eligible for reimbursement and wascategorized as a lifestyle drug. In France it was awarded 35% Social Security reimburse-ment, while in Switzerland it was registered on the reimbursement list. The EuropeanMedicines Agency is to review rimonabant following the 13 June 2007 recommenda-tion from the FDA Advisory Panel that rimonabant not be approved in the USA due toconcerns about psychiatric side-effects and the manufacturer’s decision to withdrawthe rimonabant New Drug Application in the USA.131
Rimonabant is currently undergoing assessment at NICE, and published guidance isanticipated February 2008. The remit or scope of the appraisal132 will be ti compareweight loss and maintenance of weight loss for rimonabant and standard treatments.The scope acknowledged the possibility of needing to estimate longer-term morbidityand mortality, including prevention of cardiovascular events and type-2 diabetes.
CLINICAL GUIDELINES IN TYPE-2 DIABETES WILL INCREASEMEDICATION USE
Over the coming decade, cost containment pressures will increase (Figure 1), andemerging evidence will continue to influence the pharmaco-economics of diabetes.In relation to cost-effective prescribing, the European Society of Cardiology (ESC)/European Association for the Study of Diabetes (EASD) guideline recommended7: (1)lipid-lowering treatment provides a cost-effective way of preventing complications;and (2) tight control of hypertension is cost-effective. These findings resonate withthe conclusions by Vijgen et al86 described earlier.
Multiple therapy is a reality for people with type-2 diabetes now.133 Use of pharma-ceuticals will increase as the recommendations of this guideline push toward earlierdiagnosis, more aggressive management with more patients being treated for longerto ensure that target levels of blood glucose, blood pressure and lipid control areachieved.74 This will lead to an increase in expenditure of pharmaceuticals, in the im-mediate term an increased uptake of statins, antihypertensives, and the more estab-lished oral antidiabetic agents equipped with longer-term efficacy, safety and
Pharmaco-economic issues for diabetes therapy 675
economic data, such as metformin, sulphonylureas, thiazolidinediones, and parenteralinsulin to achieve target glycaemic control.
NICE has produced guidance on the management of patients with type-2 diabetes,including the management of blood glucose levels.134 These guidelines are currentlybeing updated, and are expected to be reported in February 2008.135
SIGNIFICANT INCREASING SPEND ON MEDICINESTO TREAT DIABETES PREDICTED
US spending on diabetes treatments could soar by 70% from 2007 through 2009 dueto rising obesity, an aging population, increasing prevalence of diabetes, more aggres-sive treatment of diabetes, and a pipeline of new (and more costly) treatment op-tions.136 During 2006, diabetes treatments were the second largest contributor toprescription drug spends, trailing only cholesterol-reducing drugs. Increased usagewas largely driven by expansion of the treatment population rather than intensive ther-apy. It is expected that the new treatments for diabetes (together with novel treat-ments in the pipeline) will drive unit costs even higher as new-generation medicinesare prescribed. Increasing expenditures on pharmaceuticals can be expected in othercountries where health-care professionals aspire to achieve the targets for diabetescontrol laid out in their national diabetes guidelines.
PHARMACO-ECONOMIC ISSUES TO ADDRESSIN FUTURE APPRAISALS
Endorsement of any technology is not awarded simply because the ICER falls belowthe threshold for cost-effectiveness. Common problems identified in the pharmaco-economic literature for diabetes (and other disease areas) and HTA experiences aresummarized in Table 4.
A paucity of robust evidence on relevant longer-term outcomes measures, togetherwith a lack of HRQOL data collected in a reliable manner in appropriate patients andamenable to utility (and hence QALY) estimation, have resulted in problems. Analysesmust reflect the context of the decision-taker. Using a peer-reviewed economic modelis not sufficient; the assumptions which drive a model will be challenged to confirmtheir relevance to a specific jurisdiction. Cost pressures will increase as the evidencebase matures when phase-III and phase-IV studies for the new diabetes drugs arepublished.
GENERATING AND INTERPRETING THE RIGHTCLINICAL EVIDENCE?
To be fit for purpose, relevant RCTs available in a timely fashion and reflecting clinicalpractice must include HRQOL tools such as the EQ-5D which translate into utility andQALY gain associated with the technology in question. The RCT is the gold standardsource of quality clinical evidence137 and recommended to populate models.138 How-ever, the relevance of the RCT as a source of clinical efficacy has been challenged139,largely because the question posed is not the goal of the economic evaluation.
In a recent review of the insulin glargine and insulin detemir, it was concludedthat they both appear to demonstrate better glycaemic control than NPH insulin,with a similar or reduced risk of hypoglycaemia.140 It was also noted that insulin
676 J. M. Bottomley and F. D. Raymond
detemir appeared to be associated with less body weight increase than NPH insulinor insulin glargine. However, two reports using The Health Improvement Network(THIN) database in the UK reported that insulin glargine was more effective141 andless costly142 than insulin detemir and highlight a discrepancy between the RCT and‘real-world’ data.
The challenges in interpreting the cardio-diabetes evidence143,144 and the needs ofHTA bodies require the conduct of more relevant RCTs. Importantly, designing RCTsto meet clinical needs and reflect reality should address the scope of HTA appraisals.Aspects of study design which influence the results include patient populations, pri-mary and secondary study endpoints, inclusion and exclusion criteria, treatmentsand their duration, study locations, assumptions and definitions.
GEOGRAPHICAL AND ETHNIC GENERALIZABILITYOF COST-EFFECTIVENESS DATA
Decision-takers are aware of national variations in terms of medical resource utilizationand clinical outcomes, and are increasingly enquiring of the implications of this.145–147
Trial-wide cost-effectiveness results may not be directly applicable to any of the coun-tries collaborating in a multinational study. In addition, study variability may arise fromdifferences in baseline event rates, costs and utility.148 The importance of accountingfor between-country differences in tailoring decisions at local, regional and nationallevels is therefore a key issue to ensure that cost-effectiveness results are not dismissedas irrelevant to a particular location.
Type-2 diabetes and its complications continue to accelerate in developing coun-tries and in certain ethnic groups.6 Cardiovascular morbidity and mortality in womenis highlighted as an area meriting special attention.7 The requirements for optimaleconomic support of new technologies in diabetes must address the fact that exist-ing treatment approaches have been based largely upon trials in Caucasian popula-tions which have probably significantly underestimated the risk of cardiovascularevents.
The climate surrounding delivery of health care is changing and will continue to in-fluence pharmaco-economics in diabetes. A number of other recent developments isdescribed which could impinge upon future economic evaluations in diabetes.
A UK prescription for true value?
An independent report from the Office of Fair Trading (OFT)149 has proposed radicalreforms which if accepted could institute true value-based pricing in the UK. It has rec-ommended that the current Pharmaceutical Price Regulation Scheme (PPRS) be re-formed to deliver better value for money by being replaced with a patient-focusedvalue-based pricing system in which the prices that the NHS pays for medicines re-flects the benefits that they bring to patients. Such recommendations would fit snuglywith the NICE methodological and decision-making process. If accepted, however,these proposals would not take effect until 2010.
Health economists have recommended that NICE should cap at £20,000 per QALYrather than £30,000. MPs were also told that there was a need for routine monitoringof the impact of new treatments once recommended by NICE.150
s in diabetes.
Solution
duct of local research to benchmark the technology in
risdiction is required
rts must be made to frame the economic analysis under
ew within the context of the decision-maker or risk the
lysis being dismissed as not relevant
design is crucial to ensure that the right outcomes are
erated to investigate the right decision problem
a well-validated accepted model for lifetime simulation
e horizon of relevance in diabetes) where the model is
ulated with evidence-based and credible assumptions
ch reflect the patient pathway and locality
tomize available evidence to address decision problem
icipate decision problem and collect the required out-
es data in readiness for a submission to HTA bodies
ude the EQ-5D tool in RCTs to generate utilities and
ble the estimation of QALYs
ertainty in analyses must be assessed using the ap-
aches required of the decision-makers
ongoing review of the NICE ‘Guide to methods of
nology appraisal review process and timelines’155 will
tify best practice in the use of indirect comparisons and
ed treatment comparisons
quality adjusted life year; NICE, National Institute for Health and
Pharm
aco-eco
nom
icissu
esfo
rdiab
etesth
erapy677
Table 4. Problems and solutions in improving economic analyse
Problem Observation
Inappropriate
comparator
and or/lack of relevance
to the country or
decision-maker
� A well-designed double-blind RCT which does not have rel-
evance to a national decision problem is not fit for purpose in
HTA
� Many submissions to HTA groups are dismissed if this is not
addressed
� Con
a ju
� Effo
revi
anaOver-optimistic
assumptions� Cost-effectiveness results differ depending on different as-
sumptions used in modelling
� Most published analyses report favourable ICERs with studies
funded by industry more likely to report ICERs below cost-
effectiveness thresholds154
� RCT
gen
� Use
(tim
pop
whiInappropriate use
of clinical data� RCTs do not help answer the questions of interest in eco-
nomic evaluations
� The aim of phase-3 trials is often to demonstrate non-infe-
riority between two active treatments e an objective which
is at odds with the aim of economic evaluation to show su-
periority of one technology over the other
� Cus
� Ant
com
Non-QALY outcomes � Such outcome measures are dismissed by HTA bodies as they
do not facilitate comparison between different classes of drug
or different therapeutic areas
� Incl
ena
Uncertainty not tested
appropriately� There are numerous examples of wide variation around an
ICER leading to technologies not being approved by HTA
groups until such uncertainty is characterized or reduced
� Unc
pro
Use of indirect
comparisons� In the absence of an appropriate head-to-head RCTreporting
the effectiveness of the technology under review versus all
the commonly used comparators in a specific jurisdiction, an
indirect comparison approach has to be adopted to compare
the new technology to the other treatments
� Indirect comparisons across studies should be interpreted
with caution as they are no substitute for direct head-to-head
Health-care costs are spiralling in every health-care environment. Many countries willbe unable to afford any or some of the new diabetes medicines. Different health-carejurisdictions are using a variety of approaches to dampen down demand.151–153 Suchimplementing outcome guarantees for medicines is not new and has also been appliedto antidiabetic drugs in the USA.
Risk-sharing schemes will probably become more common. Companies would berewarded for the ongoing performance of their product. With increasingly sophisti-cated electronic records and access to primary care databases enabling effectivenessof treatments to be monitored in clinical practice, the authors predict that payersof health care will in future adjust reimbursement levels of treatments accordingly.Such progress could herald a new business model for successful pharmaceutical andbiopharmaceutical industries that deliver medicines demonstrating tangible benefitsand addressing unmet medical needs while demonstrating by guarantee their confi-dence in the outcomes associated with their technology.
SUMMARY
The last decade has witnessed a markedly increased prevalence of diabetes with in-creased associated costs. Many trials with different endpoints complicate comparisonsbetween therapies. Pharmaco-economics is now at the leading edge in terms of secur-ing the most efficient use of precious health-care resources in diabetes. Recent relevanttrends report rapid growth of national drugs expenditure, an environment in which thecost control mechanisms are anticipated to intensify, and where scrutiny over the cost-effectiveness of new technologies is increasing. The value of economic analyses willincrease only if they approximate more closely the setting in which the interventionis used in clinical practice. Failing to deliver an economic analysis fit for purpose willresult in access of new treatment options in diabetes being restricted and usage limitedto populations that have previously failed well-established and inexpensive therapies.
ACKNOWLEDGEMENTS
The authors thank Beata Kloska BSc (Hons) MSc, MCLIP, Information Specialist, Li-brary Services, Royal Society of Medicine, London, W1G 0AE, UK for her assistancewith literature searches.
CONFLICT OF INTEREST STATEMENT
The authors have no conflicts of interest that are directly relevant to the content ofthis article.
REFERENCES
1. Fox-Rushby J & Cairns J (eds.). Economic Evaluation. Open University Press, 2005.
2. Weinstein MC, O’Brien B, Hornberger J et al. Principles of good practice for decision analytic
modeling in health-care evaluation: report of the ISPOR Task Force on Good Research Practices –
modeling Studies. Value in Health 2003; 6: 9–17.
Pharmaco-economic issues for diabetes therapy 679
*3. Guidelines for computer modeling of diabetes and its complications. American Diabetes Association
Consensus Panel. Diabetes Care 2004; 27(9): 2262–2265.
4. Wild S, Roglic G, Green A et al. Global prevalence of diabetes estimates for the year 2000 and pro-
jections for 2030. Diabetes Care 2004; 27(5): 1047–1053.
5. Centers for Disease Control and Prevention. National Diabetes Surveillance System. Prevalence of
Diabetes. Accessed 9 June 2007, at http://www.cdc.gov/diabetes/statistics/prev/national/tnumage.htm.
6. Centers for Disease Control and Prevention. National Diabetes Fact Sheet: general information and
national estimates on diabetes in the United States, 2005. Atlanta, GA: US. Department of Health
and Human Services, center for Disease Control and Prevention, 2005: Available at http://
www.cdc.gov/diabetes/statistics. Accessed May 2007.
*7. The task force on diabetes and cardiovascular diseases of the European Society of Cardiology (ESC)
and the European Association for the Study of Diabetes (EASD). Guidelines on diabetes, pre-diabe-
tes and cardiovascular diseases: Full text. www.escardio.org and www.easd.org Accessed May 2007.