Mobiele menu

When outcome is a balance. Methods to measure combined utility in obstetrics. Additional study to study 556-558;560-563 all involving choosing between active treatment (ususally including cesarean section) and watchful waiting, or a analogue dilemma.

Projectomschrijving

In de verloskunde heeft medisch handelen meestal gevolgen voor zowel de moeder als de (ongeboren) baby. Dat maakt het moeilijk om aan te geven of de ene handeling ‘beter’ is dan de anderen. De baby vervroegd geboren laten worden als bij de moeder de bloeddruk sterk oploopt, bijvoorbeeld, is gunstig voor de moeder maar brengt risico’s met zich mee voor het kind. In Rotterdam is onderzocht welke methoden bruikbaar zijn om de beslissing tot handelen in dergelijke situaties te ondersteunen en om uitkomsten van onderzoek op dit gebied goed te kunnen beoordelen. Gebleken is dat het beschrijven van de mogelijke uitkomsten in ‘vignetten’, situaties die kunnen voorkomen na het wel of niet medisch handelen, hierbij behulpzaam kan zijn. Patiënten, leken en medische experts kunnen dergelijke vignetten voorzien van ‘rapportcijfers’ op grond waarvan aan te geven is welke uitkomst van het medisch handelen de voorkeur heeft boven andere uitkomsten.

Verslagen


Samenvatting van de aanvraag

Many clinical studies in obstetrics and fertility often face a multidimensional outcome, resulting in complex decisions on the design, analysis and derivation of recommendations. Two prototypical outcome problems are at the core of this HTA-study. P1: when problems arise during third trimester pregnancy (e.g. pre-eclampsia, pROM, suspected growth retardation; see also proposals 4556-4558, 4560-4563) the obstetrician has to choose between a watchful waiting versus an active (induction of labour, caesarean sectio [CS]) strategy. Active policy: less risk of intrauterine death, more risk on prematurity, labour complications, disadvantages of a CS; watchful waiting involves the reverse. Probabilities of these events range from rare to universal, and the weights attached to these outcomes show a broad range. P2: in vitro fertilisation involves a decision on the strategy to maximize success (per cycle, and overall given estimated total time) and to minimize twin (and higher) rate. Increasing the number of embryos implanted enhances pregnancy rate at a price of more multiple pregnancies, with associated risks of short term intrauterine death, prematurity, IUGR, TTS, etc and long term sequelae of prematurity (with IUGR). Clinical studies in this area typically face problems in definining of a testable hypothesis, the primay/secundary outcome measures, the sample size and finding a method to arrive at a data-supported decision on what strategy is best as is testified by many research proposals in previous ZONMW programs. Also standard QALY-calculations do not apply. We propose to compare head-to-head methodologies in use in socalled preference and utility measurement to arrive at a prefered method to measure and analyse these multidimensional outcomes. The primary comparison will be the 'attitude' method based in psychology, applying, for short rating scales and sophisticated analysis to cover personal effects and error, and the 'preference' method based in economics, with 2 branches: indirect trade-offs with an artificial calibration scale as in time-trade-off or WTP, or socalled direct discrete choice experimentation (DCE) choosing between 2 vignettes, as in conjoint analysis). This may be called the choice on the philosophy of outcome measurement (Phillips, 2002). This HTA-proposal is restricted to problem 1 (obstetrics), but if granted, it will be combined with the VIDI-project of Mol to include the parallel infertility problem, approached in an identical way. Conform Payne (1999) the proposal distinguishes the following 5 stages (innovations are present in stage 2,3,5). Stage 1:Empirical data on the outcome (mean, variance, range) will be derived from literature (Cochrane reviews, some recent UK studies), published and on going research (e.g. home surveillance of high risk pregnancies, PETRA-study), the studies from the 'consortium' (556-558, 560-562). A limited number of domains (4-6) will be selected, based on existing literature, research proposals, own research, and interviews with patients (already involved in primary research) and doctors. Stage 2:Using stage 1 data, an average case willl be composed and projected into a vignette. The format will be conform the recently published study on the quantification of mild diseases [MiDAS; Bonsel, 2003] including graphical support. Several variants of risk presentation (framing variants) will be compared (Edwards 2001). Stage 3:Selection of respons method; for the attitude approach we choose tha VAS, with indirect trade-off we choose TTO and WTP, with the direct method no choice is involved as only preference among 2 vignettes is asked for. Stage 4: Selection of judging respondent. We will use patients participating in the previously mentioned studies, caregivers (obstetricians, midwives) and laymen (a panel from the MiDAS-study is still available). Stage 5. Analysis will follow rules applicable to the 3 methods. Results will be compared on technical aspects (e.g. the consistency of ratings, the additivity and independence assumption underlying conjoint analysis, etc.) and on 'user validity' by a panel of potential users: obstetricians, pregnant women and epidemiologists. Power/data analysis is not applicable; panel samples will be in the order of 50 (patients), 100-150 (laymen), and 25 professionals depending on opportunities. If possible recommendations extend to hypothesis/sample size consequences. In the optimal case delivrables are: a best practice description for the rating and conjoint analysis method; order of magnitude estimates on the relative value of the components of the abovedescribed clinical problem; recommendations for design/hypothesis and overall outcome evaluation in these cases. It can be anticipated these delivrables have general application.

Kenmerken

Projectnummer:
945045589
Looptijd: 100%
Looptijd: 100 %
2004
2009
Onderdeel van programma:
Gerelateerde subsidieronde:
Projectleider en penvoerder:
Prof. dr. G.J. Bonsel
Verantwoordelijke organisatie:
Erasmus MC