In their paper in this issue of Addiction, Dietze et al. reported the life satisfaction of injecting drug users compared to the general population . Life satisfaction was measured using the Personal Wellbeing Index (PWI), which has seven satisfaction domains: standard of living, health, achievements in life, personal relationships, personal safety, feeling a part of the community and future security. All PWI domains were significantly lower in the injecting drug users than the general population. They also showed that increasing injection frequency and recent serious mental health problems predicted lower life satisfaction and being employed predicted higher life satisfaction across all domains. The authors suggested that results from this measure could inform more tailored treatment approaches for substance use disorder patients.
A variety of life satisfaction (or subjective wellbeing) and health-related quality of life (HRQL) measures are available. For clarity, I will focus on the PWI for life satisfaction and preference-weighted HRQL measures (e.g. EuroQol, Health Utility Index, SF-6D, Quality of Well-Being, Assessment of Quality of Life). Recent evidence supports the construct validity of preference-weighted HRQL measures in SUD patients [2–4].
There is a vigorous debate in the literature about the relative merits of subjective versus objective wellbeing for informing allocation of healthcare resources [5–8]. A major weakness cited for preference-weighted HRQL measures is that individuals assign preference weights to health states that they have not experienced, while a major weakness cited for life satisfaction measures is their strong association with personality traits, mood state, or both. An implication for life satisfaction measures informing health-care resource allocation decisions, and being correlated strongly with mood state is a greater proportion of health-care resources being allocated for mental health services. As a psychiatrist, perhaps this is not a major weakness after all.
Whose perspective to use for assigning value to quality of life is also part of the debate. For the PWI, this issue is straightforward: the patient answers the question about their own life satisfaction on a 0–10 scale. A method for assigning relative weights to various health and non-health factors affecting life satisfaction has been suggested . Aside from including non-health factors as covariates and using a continuum of death to full life satisfaction instead of death to perfect health, the suggested method resembles the one used to develop preference weights for the Quality of Well-Being (QWB) measure .
For HRQL measures, the general public is recommended typically as the source of preference weights because general public resources pay for health care, the general public is blind to self-interest and outcome valuations should be representative of the entire at-risk population. Interestingly, for most illnesses, patients assign higher preference weights (closer to perfect health) to their problem than does the general public. Explanations for this finding include scale recalibration and adaptation to the illness . One exception, however, may be depression. Depressed patients assign lower preference weights to depression health states than the general public . A possible explanation for this finding is the general public undervaluing the impact of depression due to public stigma. For depression interventions, using depressed patient preference weights would result in a bigger bang for the health-care buck than using general public preference weights. Data are not yet available, but patient preference weights for outcomes associated with substance use disorders and other mental illnesses may follow a similar pattern.
There is a low correlation between HRQL and life satisfaction measures . One explanation for this finding is relative stability of life satisfaction over time, which is described by the Hedonic Treadmill theory and the Theory of Subjective Wellbeing Homeostasis [13,14]. Other explanations for the relative stability of life satisfaction over time include the influence of personality, temperament, defense mechanisms, adaptation and coping mechanisms . Genetic and heart rate variability biomarkers are also being explored to explain individual variability in life satisfaction [16,17]. I am not aware of studies examining the association between biomarkers and the assignment of preference weights.
Stepping back from the methodological debate, I am struck by a more fundamental question. What should be the goal of health care? Is it to increase life expectancy, decrease symptoms and reduce functional impairment, or is it to maximize a broad definition of life satisfaction? If the latter, more discussion is needed about how to change health-care delivery to accommodate life satisfaction as a primary outcome. Other directions for future research include the following. First, patient-centered outcomes are important and require more attention in resource allocation decisions. This may be especially true for patients with substance use disorders, where the argument for investing in treatment has focused historically upon cost savings rather than improving quality of life. Secondly, future work should continue to identify biomarkers associated with outcomes that matter to patients.