TOG
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

The Obstetrician & Gynaecologist 2007;9:3:181-187
doi: 10.1576/toag.9.3.181.27339
Copyright © 2007 by the Royal College of Obstetricians and Gynaecologists.
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Abedin, P.
Right arrow Articles by Khan, K. S
Right arrow Search for Related Content
PubMed
Right arrow Articles by Abedin, P.
Right arrow Articles by Khan, K. S

Education

Health technology assessment in obstetrics and gynaecology. Part 2: application in practice

Parveen Abedin, MSc MRCOG, Locum Consultant Obstetrician and Gynaecologist1, Jane Daniels, BSc MMedSci MSc, Senior Research Fellow and Unit Co-ordinator2 and Khalid S Khan, MMed MRCOG, Professor of Obstetrics, Gynaecology and Clinical Epidemiology3

1. Good Hope Hospital, Rectory Road, Sutton Coldfield, Birmingham B75 7RR, UK Email: Pabedin{at}doctors.org.uk (corresponding author)
2. Birmingham Women's Hospital, UK
3. Birmingham Women's Hospital, UK


   Abstract
TOP
 Abstract
Introduction
 Routine antenatal anti-D...
 Selection of women for...
 Conclusion
 References
 
Key content:

Learning objectives:

Ethical issues:

Please cite this article as: Abedin P, Daniels J, Khan KS. Health technology assessment in obstetrics and gynaecology. Part 2: application in practice. The Obstetrician & Gynaecologist 2007;9:181–187.

Keywords anti-D prophylaxis / health technology assessment / hysterectomy / randomised controlled trial / systematic review


   Introduction
TOP
 Abstract
 Introduction
Routine antenatal anti-D...
 Selection of women for...
 Conclusion
 References
 
In Part 1 of this article an overview was presented of the process of health technology assessment (HTA). Its use by agencies such as the National Institute for Health and Clinical Excellence (NICE) in deciding on the most clinically effective and cost efficient technology for the National Health Service (NHS) was discussed. Information was given on systematic reviews, randomised controlled trials (RCTs) and economic evaluations. Where, though, does all this fit into our specialty?

As clinicians, we are faced every day with making decisions about the optimal care for women with regard to effectiveness and cost. HTAs are increasingly being used to incorporate the most clinically and cost-effective evidence-based methods into our clinical practice. An example is the increasing use of early pregnancy assessment units to diagnose early pregnancy problems, such as miscarriage and ectopic pregnancy. Women spend a minimum of time as inpatients, thus cutting down on ‘hotel’ costs, while being managed primarily by advanced nurse practitioners with access to medical expertise if needed. Another example is the community antenatal care of low-risk women. Hospital and travel costs to women are cut down, while there is continuity of care from community midwives.

The process of HTA in obstetrics and gynaecology is demonstrated by the examples given here.


   Routine antenatal anti-D prophylaxis for rhesus-negative women
TOP
 Abstract
 Introduction
 Routine antenatal anti-D...
Selection of women for...
 Conclusion
 References
 
Our first example is the HTA of routine antenatal anti-D prophylaxis given to rhesus (Rh) negative women.1 Haemolytic disease of the newborn results from the transplacental passage of maternal alloantibodies directed against fetal red cell antigens inherited from the father. Ninety percent of all cases of clinically significant haemolytic disease of the newborn affect Rh-positive infants born to Rh-negative mothers. Approximately 16% of women in the UK are Rh-negative and in about 10% of pregnancies the mother is Rh-negative and the fetus Rh-positive. Currently, it is estimated that sensitisation occurs in 625 Rh-negative women per year, leading to 30 fetal or neonatal losses. The proposed service evaluated in the HTA report is the routine offering of anti-D prophylaxis to all pregnant women who are Rh-negative or to Rh-negative primigravid women only. Intramuscular anti-D immunoglobulin would be given at 28 and 34 weeks of gestation.

Search strategy
All the major databases were searched, including MEDLINE, EMBASE, Best Evidence, CINAHL, the Health Management Information Consortium (HMIC), Science Citation Index, Cochrane Database of Systematic Reviews (CDSR), Cochrane Controlled Trials Register (CCTR) and the Centre for Reviews and Dissemination (CRD) databases (Database of Abstracts of Reviews of Effects [DARE], the NHS Economic Evaluation Database [NHS EED] and the Health Economic Evaluation Database [HEED]). They were also searched for current research and grey literature. See Box 1 for inclusion and exclusion criteria.


Figure 1
View larger version (29K):
[in this window]
[in a new window]

  		Box 1
Inclusion and exclusion criteria for a systematic review of anti-D prophylaxis for Rh-negative pregnant women

 
Quality assessment
Conventional quality assessment scales were not used in the review as there was insufficient detail to discriminate between the studies.2 The groups in the studies were assessed for similarity at baseline for prognostic factors. The reason stated for this was that there were differences in study design. This would potentially have added bias to the results. One way of avoiding bias would be to look at the RCTs as one group in a subgroup analysis. This could then be checked for quality assessment with regard to the adequacy of randomisation and intention-to-treat analysis. As the outcome was to look at sensitisation rates, which is an objective measure, blinding is not as important here as it would be in a study where, for example, the outcome is to look at pain scores or satisfaction rates among participants. Results from this group would be more reliable than from groups of other study designs.

Studies identified
A total of 11 studies were identified, which consisted of one RCT, one quasi-RCT, two before-and-after studies, six non-randomised trials with historical or geographical controls and one retrospective study. They varied in terms of patient selection and dosage regimens. Randomised controlled studies are ideal for intervention studies, as long as they are robust in their conduct and design. The authors, however, felt the need to include non-randomised studies as the dose used in the only proper RCT was lower than the one currently felt to be appropriate. It was found that the proportion of women sensitised in the intervention arm was lower than in the control arm in all of the studies. The difference varied between studies and this can be attributed to the dose or the schedule used. In two of the studies, 34 weeks of gestation was too late for routine prophylaxis.

Meta-analysis
As mentioned in Part 1 of this paper, meta-analysis should only be done if the studies in a systematic review are sufficiently homogenous. There are various statistical tests of heterogeneity and these should be used as a yardstick to ascertain whether meta-analysis should be carried out. As a result of the dosage schedule variation, the studies were divided into three groups: in group 1, a dosage regimen of anti-D 500 iu at 28 weeks and 34 weeks of gestation was used in primigravidae; in group 2, 1 500 iu at 28 weeks of gestation was used in primigravidae as well as multigravidae; and in group 3, 500 iu was used at 28 weeks of gestation in primigravidae. Tests for heterogeneity showed that there were non-significant differences among the three groups and, therefore, a meta-analysis was undertaken (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1
Results of the meta-analysis

 
Pooling the data from the 11 studies, 0.5% of women became sensitised in the intervention arms, compared with 1.4% in the control arms. The number needed to treat (the number of women who needed to be treated for one to get benefit) to avoid sensitisation was 278.

Economic analysis
An evaluation of the costs and cost effectiveness relevant to the NHS of providing routine anti-D prophylaxis to Rh-negative pregnant women was done. Before starting such an analysis, it is important to define meaningful endpoints and then decide, after reviewing the available literature, whether modelling techniques are necessary. The endpoints assessed in this evaluation were:

Eight previous economic evaluations were found after searching the literature.3–9 These included cost–benefit analysis, cost-effectiveness studies and studies using decision analytic modelling. A cost-effectiveness evaluation was then modelled of offering routine anti-D prophylaxis to all pregnant women who are Rh-negative and to primigravid Rh-negative pregnant women, in addition to the conventional antenatal anti-D prophylaxis used by the NHS.

It is important to consider all costs incurred in a cost-effectiveness study. The main considerations here are the costs incurred through providing anti-D immunoglobulin and administering the treatment (Box 2). As this study was undertaken from an NHS perspective, costs to women of travel and taking time off work were not accounted for. The economic benefits of the programme are the direct savings as a result of avoidance of sensitisation in future pregnancies.


Figure 2
View larger version (40K):
[in this window]
[in a new window]

  		Box 2
Baseline parameters for the evaluation of cost effectiveness

 
The authors estimated that the incremental cost per prevented case of RhD haemolytic disease of the newborn of the 2 x 500 dose has a central estimate of £15,241 when treating primigravidae and £48,225 when all Rh-negative women are treated. For the 2 x 1 250 iu dose, the incremental cost per case prevented was estimated to be £13,618 for primigravidae and £43,407 for all Rh-negative women.

The gross annual cost of offering routine anti-D prophylaxis to all Rh-negative pregnant women in England and Wales is estimated to be £6.1 million for the 2x1 250 iu regimen and £6.8 million for the 2x500 iu regimen. If routine anti-D prophylaxis were to be offered to Rh-negative primigravid women only, the cost of drugs would be approximately £2.1 million for the 2x1 250 iu regimen and £2.4 million for the 2x500 iu regimen.

The cost per QALY was then calculated on the basis of the published literature relating to the impact on quality of life of long-term neurodevelopmental problems in low birthweight infants. It was found that routine anti-D prophylaxis given to all pregnant Rh-negative women is economically beneficial, with a maximum cost-effectiveness ratio of £30,000 per QALY, if the loss of a child, associated parental grief and subsequent high intervention pregnancies are valued at more than nine QALYs.

Routine anti-D prophylaxis has now been introduced as standard practice in the NHS.11


   Selection of women for laparoscopically assisted hysterectomy
TOP
 Abstract
 Introduction
 Routine antenatal anti-D...
 Selection of women for...
Conclusion
 References
 
The second HTA is from the field of gynaecology. The example chosen deals with the route of hysterectomy. Generally, women who undergo abdominal hysterectomy spend at least 4–5 days in hospital and have a slightly longer recovery period than those undergoing vaginal hysterectomy. However, it may not be technically possible to carry out vaginal hysterectomy on all women and this is where laparoscopically assisted vaginal hysterectomy has a place.

The EVALUATE trial12 was designed to test the null hypothesis of no significant difference between laparoscopic and abdominal hysterectomy and between laparoscopic and vaginal hysterectomy with regard to each of the outcome measures. An economic evaluation was also carried out regarding the costs and cost effectiveness to the NHS.

Trial design
Two parallel, concurrently conducted, multi-centre randomised trials were designed. In one, the ‘abdominal’ trial, laparoscopic hysterectomy was compared with abdominal hysterectomy. In the other, the ‘vaginal’ trial, laparoscopic hysterectomy was compared with vaginal hysterectomy.

There were broad inclusion criteria for eligibility for randomisation (see Box 3). However, the exclusion criteria listed considerably narrowed the conclusions that can be drawn from the study, as many cases that would normally warrant surgery (such as a bulky fibroid uterus) were excluded. This calls into question the external validity of the study. An added source of selection bias was introduced by consultants being able to decide ultimately whether women were unsuitable for inclusion.


Figure 3
View larger version (37K):
[in this window]
[in a new window]

  		Box 3
Inclusion criteria for eligibility for randomisation

 
Randomisation
Women were selected for inclusion to the vaginal or abdominal trial as decided by the treating gynaecologist. Computer-generated randomisation was done. Imbalanced randomisation was chosen in favour of laparoscopic hysterectomy in a ratio of 1:2 (Figure 1). This was to allow for the 2–3% conversion rate from laparoscopic to open surgery and to ensure that surgeons, if not already at the top of their learning curve, gained as much experience of laparoscopic surgery as quickly as possible.


Figure 1
View larger version (24K):
[in this window]
[in a new window]

  		Figure 1
Imbalanced randomisation for the study12

 
Trial objectives
The primary objective was to record the number of occurrences of a major complication or death. The following were counted as major complications:

The following parameters were determined:

Benign conditions such as menorrhagia can be distressing and disabling. The key objective of treatment is to improve quality of life. Any trial to discover the effect of an intervention should focus on assessing improvement of the original condition. In our opinion, the primary outcome in this study should be health status, rather than complications. If complications occur they reflect on health status measurement, thereby having an impact on the outcome of hysterectomy. The occurrence of any major complication is a composite outcome measure that is usually viewed with scepticism and suspicion in clinical trials, as it invalidates inferences about effectiveness.13 This composite outcome measure has another key problem: it cannot be symmetrically applied to the various trial arms. For example, abdominal hysterectomy cannot have unintended laparotomy as a complication. If laparotomy were to be considered an undesirable outcome, all of the women allocated to abdominal hysterectomy would inevitably be recorded as having suffered this outcome. Conversion to laparotomy in laparoscopic hysterectomy can only be considered a surgical complication in light of the reason for conversion (for example, uncontrollable bleeding). The woman should be informed and have consented to the possibility of such unintended open surgery being a necessary part of laparoscopy. These concerns have been highlighted in a number of published commentaries and need to be considered in our interpretation.14

It is not uncommon for good trials comparing laparoscopic with open procedures to report data on a primary outcome of key interest to wellbeing and complication rates as secondary outcomes. For example, a trial compared laparoscopically assisted versus open colectomy for colon cancer, with the time to tumour recurrence as the primary outcome measure.15 The recurrence rate was found to be similar between the two study groups. Here, there is clear information on effectiveness.

Statistical considerations
Sample size
In any study of this kind, it is important to establish the sample size required to detect a statistically significant difference at the protocol stage. In the abdominal versus laparoscopic hysterectomy arm of the trial, it was expected that 9% of women undergoing abdominal hysterectomy would have major complications. If an incidence of 4.5% of complications was to be observed in the laparoscopic hysterectomy group, a difference of 50% between the two groups, a clinically relevant outcome would be demonstrated. To do this, a total sample size of 487 women per trial arm would be needed, using 80% power and a two-sided type I error rate of 5%. Similarly, a complication rate of 4% was expected in the vaginal hysterectomy arm of the trial. If the laparoscopic hysterectomy complication rate was 2%, a difference of 50% between the two groups would be detected: again, a clinically relevant difference. To detect this difference, a sample size of 1 141 women per trial arm was required, using 80% power and a 5% error rate.

Keeping the imbalanced randomisation rate of 2:1 (laparoscopic: abdominal hysterectomy and laparoscopic: vaginal hysterectomy) for reasons mentioned above, the number of women undergoing hysterectomy was planned at 1 800: 1 048 for the abdominal and 752 for the vaginal trial. In the abdominal arm of the trial, it was planned that 349 women would have abdominal hysterectomy and 699 assisted laparoscopic hysterectomy. Out of the women in the vaginal trial, it was planned to recruit 501 to the vaginal laparoscopic hysterectomy arm and 251 to the vaginal hysterectomy arm of the trial. In fact, only 1 380 women were recruited in total: 292 to abdominal hysterectomy, 584 to assisted laparoscopic hysterectomy, 168 to vaginal hysterectomy and 336 to vaginal laparoscopic hysterectomy. Fifty-one surgeons had agreed to participate in the trial but only 43 surgeons actually recruited women.

Intention-to-treat analysis
One of the methods of avoiding attrition bias is to analyse patient data according to the type of operation to which they were initially randomised, irrespective of whether they underwent it or not. This was done in the trial.

Results
(See Table 2)


View this table:
[in this window]
[in a new window]

  		Table 2
Major complications of hysterectomy by route12

 
Quality of life results
Quality of life (QOL) was measured using the SF-12 Health Survey Questionnaire, Body Image Scale (BIS) questionnaire, Sexual Activity Questionnaire (SAQ) and the EuroQoL instrument (EQ-5D).

A highly significant statistical difference was found in the SF-12 physical component summary score, the BIS score, the SAQ score and the EQ-5D score between abdominal hysterectomy and laparoscopically assisted vaginal hysterectomy, in favour of the latter.

Major complications
The assisted laparoscopic hysterectomy procedure had a statistically significantly higher major complication rate than abdominal hysterectomy. There was no significant difference in the rate of major complications between the vaginal and vaginal laparoscopic hysterectomy procedures. Major complications tended to happen before discharge from hospital.

A statistical method of analysis called logistic regression is used to identify any variables that can influence the outcome of a trial. Logistic regression was used here to identify variables that can influence the likelihood of a major complication.

Abdominal hysterectomy
The variables thought to be important in the prediction of a major complication in abdominal hysterectomy were:

However, none of these variables were found to be important predictors.

Vaginal hysterectomy
The following were thought to be important in influencing the rate of major complications in women having vaginal hysterectomy:

However, only previous pelvic surgery was found to influence the rate.

Laparoscopic hysterectomy
The variables thought to influence the major complication rates were:

A non-significantly higher rate of complications occurred in procedures where the larger 12 mm trocar was used. The bipolar method of securing haemostasis on the ovarian pedicle resulted in 8.6% of women having major complications, compared with 35.3% of women where the suturing method was used. This accounted for 25% of all major complications for laparoscopic hysterectomy. Uterine descent and haemostasis of both the uterine and ovarian pedicles are significantly important variables in the prediction of major complications.

Economic evaluation
A cost-effectiveness analysis was carried out in which the various routes of hysterectomy were compared. In the comparison of laparoscopic and vaginal hysterectomy, the mean cost per woman (£401) and QALY (0.0015) was higher in the former. The incremental cost-effectiveness ratio (ICER), the mean difference in costs divided by the mean difference in QALYs, was £267,333.

When laparoscopic and abdominal hysterectomy were compared, the laparoscopic procedure had a higher mean cost (£186) and higher mean QALY per woman (0.007). This generated an ICER of £26,571. Compared with vaginal and abdominal procedures, laparoscopic hysterectomy resulted in a higher mean time in the operating theatre and more extensive use of disposable instruments. This was, however, offset by a shorter mean hospital stay compared with abdominal hysterectomy.

Ultimately, it is the willingness of the NHS to pay the extra cost per additional QALY that will decide whether the laparoscopic route is adopted for hysterectomy. In terms of QALYs, there was little difference between laparoscopic and abdominal or vaginal hysterectomy. Vaginal hysterectomy appears to be more cost effective than laparoscopic hysterectomy, but cost effectiveness of the laparoscopic approach relative to abdominal hysterectomy was finely balanced. This HTA reminds us that hysterectomy complications are common, whatever the route. As such, training is an important issue, as the higher complication rate observed with laparoscopic hysterectomy can be considered to be the result of a learning curve effect.


   Conclusion
TOP
 Abstract
 Introduction
 Routine antenatal anti-D...
 Selection of women for...
 Conclusion
References
 
Health technology assessment, such as in the two examples illustrated, is playing an increasing role in ensuring that evidence-based methods dictate clinical practice. HTAs of topics of relevance are commissioned by governmental bodies, such as NICE, to direct health policy makers in approving and funding the most clinically effective and economically efficient health technologies. An amalgamation of all the information gathered on clinical effectiveness and cost effectiveness goes into the making of an HTA. This information is then scrutinised by policy makers as well as patient representatives and, finally, incorporated into practice.


   References
TOP
 Abstract
 Introduction
 Routine antenatal anti-D...
 Selection of women for...
 Conclusion
 References
 

  1. Chilcott J, Lloyd Jones M, Wight J, Forman K, Wray J, Beverley C, et al. A review of the clinical effectiveness and cost-effectiveness of routine anti-D prophylaxis for pregnant women who are rhesus-negative. Health Technol Assess 2003;7:iii–62.[Medline]
  2. Clark HD, Wells GA, Huet C, McAlister FA, Salmi LR, Fergusson D, et al. Assessing the quality of randomized trials: reliability of the Jadad scale. Control Clin Trials 1999;20:448–52. doi:10.1016/S0197-2456(99)00026-4[Medline]
  3. MacKenzie IZ, Bowell P, Gregory H, Pratt G, Guest C, Entwistle CC. Routine antenatal Rhesus D immunoglobulin prophylaxis: the results of a prospective 10 year study. BJOG 1999;106:492–7. doi:10.1111/j.1471-0528.1999.tb08304.x
  4. Tovey LA, Townley A, Stevenson BJ, Taverner J. The Yorkshire antenatal anti-D immunoglobulin trial in primigravidae. Lancet 1983;2:244–6. doi:10.1016/S0140-6736(83)90232-5[Medline]
  5. Selinger M. The pharmacoeconomics of antenatal prophylaxis. In: Building on Success: Antenatal Prophylaxis. London: Bio Products Laboratory; 1997. p. 6–7.
  6. Lim OW, Fleisher AA 2nd, Ziel HK. Reduction of Rh0(D) sensitization: a cost-effective analysis. Obstet Gynecol 1982;59:477–80.[Medline]
  7. Baskett TF, Parsons ML. Prevention of Rh(D) alloimmunization: a cost-benefit analysis. CMAJ 1990;142:337–9.[Abstract]
  8. Vick S, Cairns J, Urbaniak S, Whitfield C, Rafaat A. Cost-effectiveness of antenatal anti-D prophylaxis. Health Econ 1996;5:319–28. doi:10.1002/(SICI)1099-1050(199607)5:4<319::AID-HEC211>3.0.CO;2-6[Medline]
  9. Torrance GW, Zipursky A. Cost-effectiveness of antepartum prevention of Rh immunization. Clin Perinatol 1984;11:267–81.[Medline]
  10. Adams MM, Marks JS, Koplan JP. Cost implications of routine antenatal administration of Rh immune globulin. Am J Obstet Gynecol 1984;149:633–8.[Medline]
  11. National Institute for Clinical Excellence. Technology Appraisal Guidance No. 41. In: Guidance on the Use of Routine Antenatal Anti-D Prophylaxis for RhD-negative Women. London: NICE; 2002. [www.nice.org.uk/pdf/prophylaxisFinalguidance.pdf].
  12. Garry R, Fountain J, Brown J, Manca A, Mason S, Sculpher M, et al. EVALUATE hysterectomy trial: a multicentre randomised trial comparing abdominal, vaginal and laparoscopic methods of hysterectomy. Health Technol Assess 2004;8:1–154.[Medline]
  13. Montori VM, Jaeschke R, Schünemann HJ, Bhandari M, Brozek JL, Devereaux PJ, et al. Users' guide to detecting misleading claims in clinical research reports. BMJ 2004;329:1093–6. doi:10.1136/bmj.329.7474.1093[Free Full Text]
  14. Chien P, Khan K, Mol BW. How to interpret the findings of the eVALuate study. BJOG 2005;112:391–3. doi:10.1111/j.1471-0528.2005.00629.x[Medline]
  15. Neumayer L, Giobbie-Hurder A, Jonasson O, Fitzgibbons R, Jr, Dunlop D, Gibbs J, et al. Open mesh versus laparoscopic mesh repair of inguinal hernia. In: N Engl J Med. 350, 2004. p. 1819–27.doi:10.1056/NEJMoa040093.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Abedin, P.
Right arrow Articles by Khan, K. S
Right arrow Search for Related Content
PubMed
Right arrow Articles by Abedin, P.
Right arrow Articles by Khan, K. S

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS