The average percentage of GDP spent on health in OECD countries rose from 6.9% in 1960 to 8.8% in 2003. The health literature attributes this increase to three major factors (i) an increase in life expectancy; (ii) the use of advanced expensive diagnosis and treatment tools and methods; and (iii) a change in public demand and expectations, due to rising living standards. Obviously, the first two factors involve greater spending, but does the change in preferences necessarily lead to an extra burden on the national budget?
To tackle that question, it is important to have information on preferences. A useful tool for the estimation of preferences is discrete choice experiments, which gather information on stated preferences in a way that sheds light on the impact of different attributes on the overall benefit obtained from a particular good or service as well as the trade-offs between attributes. The technique involves presenting individuals with hypothetical (realistic) scenarios that vary with respect to the levels of defined attributes and asking them to make discrete choices between pairs of scenarios. The data generated by the discrete choices are then used to estimate preference equations that present estimates of the marginal utility scores of the attributes. Trade-offs between attributes can also be estimated (see Ryan and Gerard 2003 for a review of the use of discrete choice experiments in the health-care sector).
In recent research, we estimate preference patterns of Israeli women for two health-care facilities: (i) maternity-wards where women stay during labor, intra-partum care and post-natal care; and (ii) oncological institutes where women who had been diagnosed with breast-cancer receive chemotherapy . The data generated were used to examine the following questions:
- What is important for women seeking medical treatment? The attributes we ask questions about for new mothers are: professionalism of the treating doctor; attitude of staff towards patients; information transferred from staff to patient; travel time; and number of beds in hospitalization room. For the breast cancer patients we looked at: the first four attributes, but the fifth was replaced by professionalism of institute (number of beds is irrelevant for this outpatient service).
- Are preference patterns different for different medical problems? A comparison of preferences for services of maternity-wards versus oncological institutes can help answer this question. For instance, does the quality of treatment matter more to patients in life-threatening situations?
- Are medical care-givers aware of patients' preferences? Full information on the patient's preference pattern is crucial for the doctor (agent) in order to act efficiently on the patient's (principal's) behalf. Discrete choice experiments were conducted simultaneously within samples of patients (women who had given birth) and care-givers (nurses and doctors treating them) to elicit and contrast patients' authentic preferences with what care-givers believe them to be.
- Do preferences change as patients accumulate experience? We compared three sub-samples: women pregnant with their first child (no experience); women after one delivery (single experience); and women after more than one delivery (multiple experiences).
- Does the individual's socio-economic background affect preferences?
What patients care about
In our experiments, 323 new mothers made 2099 discrete choices. The data show that 'professionalism of staff' ranks first, followed by 'attitude of staff towards patients', 'information transferred from staff to patients', 'travel time from residence to hospital' and 'number of beds in room of hospitalization' (the last has an insignificant effect).
It is not surprising that 'professionalism' is the most important trait of a surgical procedure. However 'attitude' is valued only somewhat less and the most unexpected finding is that the room facilities are not valued at all. Two beds in a room and even a private room are not significantly preferred over a three-bed room. Preference equations from 95 women (1082 observations) diagnosed with breast-cancer and treated in oncological institutes show that the most important attributes are the 'professional level of the institute' and the 'reputation of the doctor'. 'Information', 'attitude' and 'travel time' come next with significant effects on utility.
Preferences are surprisingly similar
The two very different groups of patients – one experiencing the joyous medical event of a delivery and the other suffering from a life-threatening disease – exhibit quite similar preference structures. Both are willing to relate to trade-offs between attributes.
Medical staff misperceive patients’ preferences
30 doctors and nurses employed in maternity wards filled out the same questionnaires (360 observations) but were asked to make the discrete choices they believed would be made by the hospitalised women. Their answers reveal a biased perception of patients’ preferences. The staff expressed the arguably somewhat arrogant and condescending view that the type of room is the most important attribute. They thought a private room was valued well above staff professionalism. The finding that care-givers seriously underestimate patients’ valuation of ‘information’ is further evidence that doctors and nurses believe that women patients are basically ignorant.
Experience changes preferences
A comparison of preference patterns of the three sub-groups revealed strong differences between women in pre-natal classes and those who have already had a baby, but preferences did not change with more birth. Interestingly, women in pre-natal classes have preference structures that are quite similar to those expressed by care-givers. In particular, they ranked the room facilities highest. This finding confirms that care-givers attribute ignorance as to what really matters in "serious" health-care to all patients, whereas only "first-timers" exhibit a certain naivety regarding what matters.
Women with an academic education place greater value on professionalism and a private room. High-income women also place more value on a private room while women over 35 perceive themselves as benefiting more from professionalism. Within the sample of women diagnosed with breast-cancer the valuation of attitude increases with patient age and, as might be expected, academic women value shorter travel times considerably more.
These findings point to several possible policy implications that could lead to improvements in patients' satisfaction:
- An increase in patients' utility does not require major budgets and costs – factors like 'attitude of staff' and 'information transferred from staff to patients' are more important than room facilities and can be provided quite easily. All it needs is awareness on the part of the personnel and perhaps some training.
- All units of the hospital need to provide a similar policy for patient treatment and therefore care-givers who move from one unit to another do not need any adjustment – while proficiency is most valued, 'attitude' and 'information' are also highly valued by patients with very different medical problems;
- Hospital units that want to attract different populations have to apply somewhat different treatment policies. Alternatively, care-givers should be aware of these differences in order to cater better to the needs of the various socio-economic population groups.
Louviere, J., Hensher, D. and Swait, J. (2000), Stated Choice Models: Analysis and Application, Cambridge: Cambridge University Press.
Ryan, M. and Gerard, K. (2003), "Using Discrete Choice Experiments to Value Health Care Programmes: Current Practice and Future Research Reflections," Applied Health Economics and Health Policy 2(1): 55-64.
Wilcock, A., Kobayashi, L. and Murray, I. (1997), "Twenty-Five Years of Obstetric Patient Satisfaction in North America," Journal of Prenatal and Neonatal Nursing 10(4): 36-47.