Sundhedsøkonomi

Childbirth can be painful and may result in perineal tears. The norm for birthing is a “land” delivery. An alternative option is a “water” delivery. These two delivery types have different associated costs as well as different effects on perineal tears and length of labor. There is no statistical difference in child well-being outcomes as measured by Apgar scores at the first minute. The national health care system is considering offering water birthing as a treatment option to women (and therefore investing in special equipment) who will be going through labor. You are asked to help evaluate this choice.
Table 1 presents data on deliveries costs for the two procedures:
Table 1: Costs
Variables Unit Cost
(DKK, 2007)
Cardiography (per hour) 292
Standard labor assistance 2920
Standard labor assistance exceeding 2 hours (per hour) 487
Pool cleaning (water delivery) 182
Additional obstetrician during labor (per hour) (Water delivery) 487
Additional obstetrician at delivery (water delivery) 487
Suture for first degree perineal tear 244
Suture for second degree perineal tear 974
Suture for third degree perineal tear 1220
Perineal episiotomy 974
Newborn assistance (at first minute):
     Apgar score>9 0
     Apgar score = 8-9 292
     Apgar score = 7 536
     Apgar score <7 827
Newborn assistance (at 5th minute):
     Apgar score = 8 390
     Apgar score <7 2090

Data was gathered about women who had either given birth through a water or land delivery. Water deliveries require that women have expectations of a normal delivery with no complications. This was also required of the land delivery cases included. The data included women of maternal ages between 18-35, and gestational age between 37-42 weeks (40 weeks being the expected age at birth), with no complications in the newborn’s health. [Note the data was collected retrospectively]. Thus the women and the children they delivered had approximately the same physical characteristics and expectations.
Outcomes of child birth by delivery technique are reported in table 2.

Table 2: Outcomes and average costs by delivery type
Variables Land delivery Water delivery Difference
(95% confidence interval)
Outcomes
(Standard deviation)  
Labor duration (hours) 4.99 (2.45) 4.82 (2.29)
Perineal tears (proportion) 0.727 0.527 0.2 (0.19-0.21)
Apgar score at first minute 9.28 (0.98) 9.48 (0.62) 0.2 (-0.017 to 0.417)
Apgar score at fifth minute 9.84 (0.46) 9.95 (0.23) 0.11 (0.013-0.206)
Mean total costs (DKK 2007) 6490 9120 2630 (2470-2790)

Note that the 4th column of table 2 provides information about the differences between land and water deliveries that includes a 95% confidence interval. For those of you uncertain about what that means, note that it tells you whether, with 95% confidence, there is a statistical difference between the results. In particular, note that for the Apgar scores at the first minute, zero is included in the confidence interval (as it ranges from -0.017 to 0.417). Thus there is no significant difference between the babies’ scores as a function of type of delivery.

Table 3 indicates the differences in degree of severity of perineal tears/episiotomy between the two delivery methods.
Variables Land Delivery Water Delivery
All perineal tears (proportion) 0.727 0.527
1st degree tears (proportion of all) 0.25 0.34
2nd degree tears (proportion of all) 0.15 0.50
3rd degree tears (proportion of all) 0.35 0.15
Episiotomy (proportion of all) 0.25 0.01
Mean costs from perineal damage (DKK 2007) 651 398

This short term study does not analyze long term consequences of either maternal effects from perineal tears or of higher apgar scores on child outcomes after immediate care. It also does not include the capital costs of building new facilities for water deliveries. We leave these issues to a later portion of the exam. First,

Questions/tasks

1) Given the information on costs and effectiveness of each program, provide a cost-effectiveness analysis of the water delivery versus the land delivery option. Imagine that we are concerned in this study about maternal well-being only.
a. Clearly identify what the ‘no-intervention’ option is and why.
b. You will need to choose an appropriate measure of effects and determine if the total mean costs should be adjusted. Explain carefully what you do and why.
c. Describe the price-per-effect condition under which you would advise decision-makers to offer water deliveries as a choice for mothers.
d. Use the information in the 95% confidence intervals to generate a low and a high estimate for the range of possible results in your analysis. Think carefully about how you will be sure you have the lowest possible estimate and the highest possible estimate given your information. Describe the results as a sensitivity analysis. Is your recommendation in (b) changed?

2) Given the information on costs and effectiveness of each program, provide a cost-effectiveness analysis of the water delivery versus the land delivery option. Imagine that we are concerned in this study about infant well-being only.
a. Clearly identify what the ‘no-intervention’ option is and why.
b. You will need to choose an appropriate measure of effects and determine if the total mean costs should be adjusted. Explain carefully what you do and why.
c. Describe the price-per-effect condition under which you would advise decision-makers to offer water deliveries as a choice for mothers.
d. Compared to your study above when maternal health effects were the only consideration, are you more or less confident in your assessment? Why?

3) Is this a mutually exclusive or an independent set of program options? Describe in general the difference between mutually exclusive and independent programs. When do you use the average cost/effectiveness ratio and when do you use the incremental cost/effectiveness ratio? Explain.

4) Explain how a price-per-effect decision-making system works, as compared to:
a. An unlimited budget
b. A specific budget
In each case, explain the expected consequences for spending and outcomes

5) Explain two ways that you might include effects for both maternal and infant outcomes in one ratio (Identify the ways by their standard analytical name as well). What advantages and disadvantages are there to these? Be thorough and specific for this case. What, if any, additional information would you need? What, if any, additional assumptions would be required? Which of the three methods might be most appropriate in this case?

The Problem, Part 2:

Now assume that the longer run is being considered. The national health care system is considering installing pools for water births at hospitals. These pools would cost 100,000 DKK (2007) each at the present time (t=0) and would last for 30 years (to year 29) with 15,000 DKK (2007) spent in operating costs each year beginning with year t=1. If 50 pools are installed then we assume that all women wanting a water birth can select one. If 25 pools are installed then 75% of women wanting a water birth can select one. Assume that there are 1500 women who want to use water delivery each year and that the individual costs and effects each year are not expected to change in nominal terms over time and can be realised starting a year from now, in year 1 (so that they are nominally the same as in Table 2). For simplicity assume that there are no additional benefits or costs from perineal tears other than those occurring at delivery.

6). Are these programs perfectly divisible? Explain your reasoning and the consequences of your answer for analysis of the problem.

7). What is the present value of the costs of one pool and of 50 pools at a discount rate of:
a. 4%?
b. 1%?

7). Given this additional information on costs, generate a cost-effectiveness analysis that considers whether to build no pools (do nothing option), build 25 pools, or build 50 pools. Assume again that maternal well-being is the only effect that is being considered. Assume the discount rate is 4%.
a. Explain carefully what you do.
b. What should the recommended course of action be if there is a price-per-effect of
1. 15,000 DKK?
2. 20,000 DKK?
  3. 25,000 DKK?

8) Discuss the role of infant well-being in your analysis. Assume that you learn that a 10% increase in Apgar scores at 5 minutes translate into a 1% increase in IQ, and that IQ is imperfectly but positively correlated with higher lifetime annual wage earnings. Describe (without formally changing your analysis):
a. In economic terms, of what is this an example?
b. What sort of measurement this uses for valuing the problem in (a)?
c. Benefits and limitations of this measure in capturing the full value of the problem in (a).