The Estimation of Utility Weights in Cost-Utility Analysis for Mental Disorders: A Systematic Review

Abstract

Objective

To systematically review approaches and instruments used to derive utility weights in cost-utility analyses (CUAs) within the field of mental disorders and to identify factors that may have influenced the choice of the approach.

Methods

We searched the databases DARE (Database of Abstracts of Reviews of Effects), NHS EED (National Health Service Economic Evaluation Database), HTA (Health Technology Assessment), and PubMed for CUAs. Studies were included if they were full economic evaluations and reported quality-adjusted life-years as the health outcome. Study characteristics and instruments used to estimate utility weights were described and a logistic regression analysis was conducted to identify factors associated with the choice of either the direct (e.g. standard gamble) or the preference-based measure (PBM) approach (e.g. EQ-5D).

Results

We identified 227 CUAs with a maximum in 2009, 2010, and 2012. Most CUAs were conducted in depression, dementia, or psychosis, and came from the US or the UK, with the EQ-5D being the most frequently used instrument. The application of the direct approach was significantly associated with depression, psychosis, and model-based studies. The PBM approach was more likely to be used in recent studies, dementia, Europe, and empirical studies. Utility weights used in model-based studies were derived from only a small number of studies.

Limitations

We only searched four databases and did not evaluate the quality of the included studies.

Conclusions

Direct instruments and PBMs are used to elicit utility weights in CUAs with different frequencies regarding study type, mental disorder, and country.

Appendices

Appendix 1

See Table 4.

Table 4 Characteristics of preference-based instruments (selection based on its frequency)

1.1 Appendix 2: Search Strategy (PubMed)

((cost-effectiveness OR cost-utility OR cost-benefit OR “economic evaluation” OR economic*)) AND ((((((((“Mental Disorders”[Mesh])) OR (anxiety OR panic OR phobia)) OR (dement* OR Alzheimer OR “cognitive impairment” OR Parkinson)) OR (“anorexia nervosa” OR “bulimia nervosa” OR binge* OR purging OR “eating disorder*”)) OR (psychotic* OR schizophreni* OR “positive symptom*”)) OR (“somatoform disorder” OR somatiz* OR somatis* OR “conversion disorder” OR multisomatoform* OR “medically unexplained symptoms” OR “medically unexplained illness” OR psychogen* OR nonorgan* OR (psychosom* AND syndrome) OR “functional somatic syndrome“ OR “functional syndrome” OR hypochondri* OR “illness phobia” OR “health anxiety” OR “body dysmorphic disorder” OR dysmorphophobia)) OR (depression OR “depressive disorder”))) AND (“rating scale” OR “visual analogue scale” OR “vas” OR “category scale” OR “ratio scale” OR “person-trade-off” OR “person trade off” OR “person trade-off” OR “willingness-to-pay” OR “willingness to pay” OR “willingness-to-accept” OR “willingness to accept” OR “standard gamble” OR “time trade off” OR “time trade-off” OR “time tradeoff” OR “tto” OR “sf-6d” OR “sf6d” OR “eq-5d” OR “eq5d” OR “euroqol” OR “eqvas” OR “eq-vas” OR “health utility index” OR “health-utility-index” OR “hui” OR “15d” OR “quality of well-being” OR “qwb” OR “qwb-sa” OR “aqol” OR “rosser-kind” OR “rosser kind” OR “halex” OR “sf-36” OR “sf36” OR “sf12” OR “sf-12” OR “sf8” OR “sf-8” OR “whoqol” OR “who-qol” OR “whoqol-bref” OR “whoqol bref” OR “who-qol-bref” OR “who qol bref” OR “who-qol bref” OR “depression-free-day” OR “depression free day” OR “sah” OR “coop” OR “sickness impact profile” OR “sip” OR “health assessment questionnaire disability” OR “haq-di” OR “haqdi” OR “barthel-index” OR “barthel index” OR “depression free-day” OR “nottingham health profile” OR “nottingham-health-profile” OR qaly OR qalys OR “quality adjusted life year*” OR utilit* OR “contingent valuation” OR “Hamilton Depression” OR “General Health Questionnaire”).

1.2 Explanations

aqol = assessment of quality of life

• halex = health and activity limitation index

• sip = sickness impact profile

• haqdi = health assessment questionnaire disability

• sah = self-assessed health question

• coop = darthmouth primary care cooperative information project

Note: The “depression OR depressive disorder” term also encompasses mania and bipolar disorder.

Appendix 3

See Table 5.

Table 5 References to excluded articles

Appendix 4

See Table 6.

Table 6 Categorization of sources with respect to disorder and country