1. 15.03.2005Tor Iversen Lecture 8: Distributional considerations Reference on the reading list: Williams and Cookson

2. Problem to be discussed: With voluntary insurance each person decides for himself the kind of insurance contract he would like to have. Hence, the allocation of health care and the implied allocation of health is not a public concern. With compulsory insurance the allocation of health care and the distribution of health in the population become challenges for political assemblies Hence, distributional considerations (equity in health) become important to consider

3. What is meant by equity? According to Williams: reducing inequalities in health Assume that health is easily measurable, and interpersonal comparable for instance in terms of Quality Adjusted Life Years (QALYs). (Of course problematic, but convenient for sorting out various concepts at this stage)

4. Utilitarianism – maximize total health Health of A Health of B F X F F-F health frontier – the farthest you get with the available resources and the individuals’ capacity to benefit from health services in terms of health improvements. The welfare function has linear indifference curves with 45 degree angles with both axes. This implies that a health improvement is of equal social value independent of who is having it and on what level of health he is initially. 45 ◦

5. A shift in the utility frontier Health of A Health of B F Y F’F’ F-F has shifted to F’-F’ to the benefit of A We can think of two reasons: -B has a smaller capacity to benefit for instance because of old age -A’s disease has been less expensive to treat The utilitarian approach is not concerned about why the shift occurs. X F F’F’

6. Greater weight for A than B Health of A Health of B F X F Welfare contours steeper than 45 degrees For some reason A deserves a greater weight than B

7. Health of A Health of B F X F The welfare function has convex non- linear indifference curves The better health you have the smaller is the weight that an additional improvement in health is given in the social valuation

8. Health of A Health of B F X F The welfare function has non- linear kinked indifference curves Only improvement in health for the person with the lowest health level is given a positive social valuation Maxi-min Allocate resources to maximize the level of health to the person who is least well off. Need Seriousness of disease 45 ◦ Very strong preference for equality in health

9. An example (see also introduction): Prioritizing services and groups of patients according to explicit goals Related to degrees equity concern

10. Alternative rules (criteria) for making priorities: A: Priority according to the seriousness (prospects without the treatment) of the disease Prioritize according to increasing survival without treatment B: Priority according to treatment effect Prioritize according to difference in survival with treatment and without treatment, such that the group with the greatest difference is given first priority. C: Maximize total health within the resource constraint Prioritize according to increasing cost per saved life, such that the group with lowest cost per saved life is given first priority. D: Priority according to the seriousness of the disease constrained by an upper limit on cost per saved life. E: Maximize total health constrained by a lower limit on the seriousness of disease

11. The importance of criteria for prioritizing the three treatments

12. Some implications: Optimal priority-setting depends on the aims that the health sector is expected to pursue including the distributional considerations It is possible to obtain a considerable total health gain by prioritizing treatments with modest effect given that they are sufficiently inexpensive Criteria C and E is at a disadvantage for patients who, because of some reason, do not manage to get much health out of the health services The cost of treatment relative to other treatments should not influence priority according to criteria A and B. The introduction of cost saving technologies should influence priorities according to criterion C (and possibly D and E), but not according to criteria A and B Cost- benefit analysis are relevant for priority decisions only according to criteria C, D and E Hence, if you are in favor of criteria A or B, it is inconsistent simultaneously to argue that cost-benefit analysis should have an increased role as a means to allocating resources within the health sector

13. In fact the three types of criteria mirror quite well the criteria actually used in legislations and other regulation of priority-setting Norway: Act on the right of patients (1999), Article 2 (my translation): The patient has a right to necessary help from the specialist health service when: 1.The patient has a considerable reduction in length of life or quality of life if treatment is postponed. 2.The patient may have an expected health improvement from the treatment 3.The expected cost is in a reasonable relation to the expected effect of the treatment Which of the three criteria apply here? Ammendment (2003): The specialist health service should, based on medical criteria, decide on a time limit for a patient with a right to treatment …… If the regional health enterprise does not manage to give a patient necessary treatment within the prescribed time limit, then the patient has a right to receive treatment without further delay, if necessary from a private provider or a provider abroad. Tendencies in development of criteria over time?

14. In Norway the ambition of equity in the allocation of medical care is high among all political parties. The instruments for auditing whether these goals are fulfilled are not equally ambitious. Decentralized decision-making makes it hard to verify whether approved priority rules in fact are adhered to An example of a Norwegian study – see http://www.hero.uio.no/publicat/2005/HERO2005_2.pdf The Norwegian Act on Health Enterprises states that the aim of the health enterprises is to provide high quality specialist health care on an equitable basis to patients in need, irrespective of age, sex, place of residence, material resources and ethnic background. To what extent is this goal fulfilled?

15. Patient flows (system of referrals not strictly adhered to when the study was done) Patient GP Private specialist Hospital outpatient Hospital inpatient

16. Panel data of survey of living-conditions (Statistics Norway) merged with data on capacity and accessibility to general practice and specialist care A representative national sample consisting of 3501 individuals Reported use of outpatient visits and visits to private specialists during the last year Self-assessed health Accessibility index: Describing the access to specialist health care at the municipality level Specialist care includes: -hospitals, outpatient and inpatient care -privately practicing physician specialists Incorporated in the index is: -capacity of the specialized health care -discounted distance

17. Table 1: Distribution of contracts with private specialists and distribution of population according to regional health enterprise. Health East Health South Health West Health Middle Health North Sum Distribution of private specialists 48 %18 %17 %10 %6 %100 % Distribution of population36 %20 % 14 %10 %100 % Suspicion that actual use of the services is skewed as well The number of private specialists is about 10% of the total number of physician specialist

18. Private physician specialists with a contract 2003 Number of man-labor years per 1000 inhabitants Kilde: Statistics Norway

19. VariableDefinitionPrivate specialist visits N=589 (17%) Hospital outpatient visits N=667 (19%) The whole sample N=3501 Very good health=1 Very good health 0.190.17 0.26 Good health=1 Good health 0.460.41 0.50 Fair health=1 Fair health 0.230.27 0.17 Bad or very bad health=1 Bad or very bad health 0.120.15 0.07 Chronic illness=1 if the person has a chronic illness 0.460.49 0.31 Man=1 if man 0.380.39 0.48 AgeThe respondent’s age 47.1 (16.5) 48.4 (16.5) 44.6 (16.6) Junior high=1 if highest education is junior high 0.160.21 0.18 University low=1 if university degree after 3 years 0.21 University high=1 if university degree after 5 years or more 0.060.03 0.05 Household incomeHousehold income in NOK 100,000 3.86 (7.93) 3.35 (2.71) 3.38 (3.85) Descriptives

20. VariableDefinitionPrivate specialist visits N=589 (17%) Hospital outpatien t visits N=667 (19%) The whole sample N=3501 GPs per 10,000 residentsNo. of GPs per 10,000 residents in the municipality 7.71 (1.53) 7.81 (1.69) 7.79 (1.67) Personal GP=1 if the individual consider himself to have a personal GP. 0.680.69 0.60 Regular health center=1 if individual sees a regular health center 0.210.23 0.24 Hospital outpatient visits=1 if individual has been at hospital for outpatient visits during the last 12 months 0.311 0.19 Number of hospital outpatient visits The number of hospital outpatient visits during the last 12 months. 0.69 (2.16) 2.58 (6.35) 0.49 (2.95) Visits to a private specialist =1 if individual has had at least one visit to a private specialist during the last 12 months 10.27 0.17 Number of visits to private specialists The number of visits to privately specialists during the last 12 months 2.34 (4.89) 0.68 (2.33) 0.39 (2.19) Access to hospital bedsAccessibility index estimated for hospital beds 1.83 (2.43) 1.52 (2.25) 1.68 (2.37) Access to hospital physician Accessibility index estimated for hospital physicians 2.10 (2.84) 1.70 (2.58) 1.90 (2.73) Access to private specialists Accessibility index estimated for private specialists 2.70 (3.57) 2.03 (3.19) 2.31 (3.42) Descriptives II

21. Estimated effect of independent variables (Statistically significant at 5% level) on the probability of at least one visit to private specialists or hospital outpatient departments during the last 12 months. Marginal effect estimated from a multinomial logit model

22. Conclusion: A person’s self-assessed health contributes to the probability of hospital outpatient visits and visits to private specialists in the sense that poorer health increases the probability of a visit The utilization of hospital outpatient clinics is not influenced by socioeconomic factors A person’s accessibility to hospital physicians does not contribute to the probability of hospital outpatient visits The utilization of private specialists is influenced by the accessibility to specialists and a patient’s socioeconomic characteristics, such as education and income An individual with a higher university degree living in a municipality with the best access to private specialists has a 46 percentage points higher probability of at least one visit to a private specialist compared with an individual with junior high living in a municipality with poorest access to private specialists Private specialists seems to offer services that are supplemental to services provided by the outpatient departments and alternative to services offered by the GPs. Implications for health policy? Should specialists be relocated or ambitions regarding equity be reduced?