Introduction
The various combinations of strategies for managing idiopathic constipation in children are numerous, combining pharmacological treatments in various doses, switching treatments where one fails and titrating doses as treatments succeed or fail. There is a clear obligation on healthcare providers to provide treatments that are safe and effective and provide the greatest relief from suffering at the lowest possible cost since, where resources are finite, lower costs of care mean that more people can be treated for this condition or for other health problems. However, treatment with the lowest cost drug does not mean the most cost-effective treatment since the cost of failure associated with drugs that are less effective may outweigh the cost of higher priced alternatives. Furthermore, high cost drugs may be cost effective where they provide more health gain at an acceptable additional cost.
Cost-effectiveness analysis can provide insights into which treatment strategies provide the best health outcomes for the available NHS resources. Decisions on whether a more costly treatment is ‘worth’ the additional benefit are decided on the basis of additional cost per additional health gain. In order to be able to make comparisons across different health outcomes and maximise the use of NHS resources, NICE prefers health gain to be measured in terms of the quality adjusted life year (QALY) which is a generic measure of health benefit taking into account both years of life and quality of life. NICE has a guiding principle that an intervention is cost effective compared to the next best treatment if the additional cost per QALY is less than £20,000.
Health economic modelling can be helpful in developing guideline recommendations by showing the costs and benefits of all the alternative treatments available for a given population of children, including the downstream consequences of therapeutic success and failure. The economic evaluation of alternative treatments for idiopathic constipation requires data on both the costs and the consequences of using each treatment option. Although a wide range of treatments are available for disimpaction and maintenance for children and are prescribed by NHS practitioners, there is sparse clinical evidence of clinical effectiveness or of the downstream costs and consequences when treatments fail.
Review of the published economic evidence
A review of the health economics literature identified three studies by the same team of authors 63, 139, 140 addressing the cost effectiveness of polyethylene glycol (PEG) 3350 plus electrolytes for the treatment of faecal impaction in children. Some of the studies used the specific brand of PEG Movicol in the analysis rather than the generic term macrogol.
The first of these studies63 aimed to estimate the clinical and economic impact of using PEG 3350 plus electrolytes in outpatient settings compared to enemas and suppositories and manual evacuation to treat paediatric faecal impaction. This is a UK based economic analysis of treatment for childhood constipation and the data were based on clinical practice in England and Wales. A retrospective cohort study of 224 children aged 2 to 11 years with faecal impaction who initially received either PEG 3350 plus electrolytes, enemas plus suppositories, or manual evacuation alone for initial disimpaction was undertaken. The follow-up time was 3 months after disimpaction. The results showed comparable outcomes across groups. QALY values were reported but the quality of life weights were taken from previously published studies on constipation (0.94 for healthy children ages 2 to 11 years and 0.66 QALY for adults with constipation). The authors developed an algorithm to adapt the quality of life values for constipation in adults (0.66) to a value of 0.70 for children with constipation. Details of their methods are not given in the paper. The results of the analysis reported an equal number of QALYs at 3 months irrespective of treatment (0.21 in all groups, 95% CI 0.18 to 0.24), therefore a cost-minimisation analysis was sufficient. It showed that PEG 3350 plus electrolytes was the preferred option on cost alone (£694 versus £2759 for enemas and suppositories respectively and £2333 for manual evacuation).
The second study was also a UK based study140 that estimated the cost effectiveness of Macrogol versus lactulose for the treatment of chronic functional constipation in adults older than 18 years. The study enrolled 977 patients. Authors obtained quality of life weights from 308 members of the public using appropriate health economic techniques (standard gamble and time trade-off methods). The economic model limited the analysis to three-month cycles for treatment and patients were categorised as either successfully treated or not during this period. The authors concluded that Macrogol was a cost-effective option relative to lactulose, the same conclusion as the first study.
The quality of life weightings reported for this study were 0.74 (95% CI 0.71 to 0.75) for adults experiencing symptoms of constipation and 0.90 (95% CI 0.88 to 0.93) for people suffering from constipation but being well managed.
The final paper was an Australian study139 which looked at the costs and consequences of oral Macrogol in the disimpaction of paediatric faecal impaction in children aged 4 to 11 years. The model compared oral Macrogol with either enemas plus suppositories or with manual evacuation alone. Model inputs (clinical outcomes and quality of life weightings) reported in this paper were obtained from the earlier studies. The authors found that oral Macrogol was a cost-effective treatment for faecal impaction when compared to other alternatives.
Health economic analysis undertaken for the guideline
The body of published health economic evidence is sparse and does not address the scope of this guideline; therefore additional health economic analysis was required.
The aim of the health economic analysis for this guideline was to develop a model to compare all the pharmacological interventions and combinations of interventions that could be offered to a child with idiopathic constipation. The comparisons of drug therapies in the model are those the GDG considered to be widely used in practice in England and Wales rather than simply mirroring the comparative analyses in the published literature which did not reflect usual practice. The intention was to undertake a cost–utility analysis within a decision analytic framework comparing the different modalities of treating children with a history of idiopathic constipation confirmed by a first physical examination in terms of incremental cost per QALY. It became clear early on in the development of the guideline that the data on clinical effectiveness would be sparse. The health economic analysis used estimates made by the GDG since mean dosages and effect sizes for treatment were missing for almost all alternatives. We are aware of the limitations of this approach and discuss its implications for each of our analyses throughout this chapter.
The interpretation by the GDG of the clinical data on effectiveness was that there was no robust evidence of difference between pharmacological preparations used as first line treatment for disimpaction and that one strategy could not be recommended over any other on effectiveness grounds alone. Therefore recommendations for clinical practice should be based on other factors affecting concordance with treatment in children, such as tolerance and palatability, time to disimpaction in the initial phase of treatment and ease of use, as well as cost to the NHS.
The health economic analysis for this guideline was undertaken with the deductive assumption that all first line pharmacological strategies had the same level of effectiveness, although different assumptions provided by the GDG were used for some of the second and third line treatments where first line treatments failed. The decision to take this approach was made by GDG consensus given the absence of data on the comparative effectiveness of these treatments, and given that these treatments are currently used interchangeably in the NHS. Failure is defined as ongoing constipation requiring further treatment. The GDG was interested in finding out the difference in cost for a range of strategies for disimpaction and maintenance and whether the cost of a high-priced drug would be offset by the lower cost of failure if that high-priced drug was more effective, leading to overall savings. The economic analysis also compared the total costs per patient (including the cost of failure) of various pharmacological strategies, and considered the effect of different doses of treatment where these clinical data were available.
The economic analysis also calculated thresholds of cost effectiveness of treatment. Where one treatment or group of treatments was more effective than the alternative, there would need to be some additional therapeutic benefit of the more expensive option in order for it to be the preferred option on cost-effectiveness grounds. This additional therapeutic benefit was converted into quality adjusted life years in order to apply the NICE threshold of £20,000 per QALY to this analysis. Data on QALY weights were obtained from the published literature reviewed above.
The only data identified which estimated the effectiveness of different doses of treatment was one small study based on treatment with PEG 3350 plus electrolytes. An economic analysis of the cost effectiveness of treatment by dose was undertaken using this clinical effectiveness data.
Methods
i. Cost analysis of treatments for disimpaction
Different treatment pathways were modelled: treatments for disimpaction covered oral pharmacological treatments, in various preparations and dosages, as well as other methods of treatment such as suppositories, enemas and manual evacuation. Treatments for the maintenance phase once disimpaction has been achieved included lower dose pharmacological treatments as first line treatment, with higher doses, combinations of treatments and other more invasive procedures where pharmacological treatments fail.
The cost analysis was based on a hypothetical case of a constipated child age 5 years treated in a primary care setting with no indication of a serious underlying disorder after history and physical examination. The time frame is the first 3 months after first referral (disimpaction followed by maintenance up to 3 months). It was assumed that the maintenance dose was equal to half of the disimpaction dose. Equal numbers of follow-up hospitalisations and outpatient visits were considered across treatments. Four different pharmacological treatment groups were compared (see and ).
Disimpaction treatments, mean times to disimpaction, mean daily doses.
Unit costs used in the model, 2008.
For each pharmacological treatment two different starting doses were considered (lowest and highest reported on the BNF for Children (BNFC) website (last accessed December 2008). Combinations of treatments included baseline dosages for the different options (). The pathways for such doses are summarised in .
Cost analysis of disimpaction treatments: treatment pathways. P = probability of having success Combination = combination of treatments
The exercise was repeated using three different rates of success: low (20% success rate); medium (50% success rate); and high (80% success rate). In total, 21 different pathways were modelled.
Resources use was calculated for each pathway, including pharmacological treatment costs and hospitalisation costs (related to manual evaluation and enemas only). Data sources for unit costs are summarised in and unit costs used in the model reported in . Days of hospitalisation for enemas and manual evacuation were assumed to be 4 days in the first instance, and sensitivity analysis was performed to assess the impact of fewer days of hospitalisation required (day case, 2 days, 3 days). In reality, children come into hospital as a day case for manual evacuation and don't come in at all for the enemas. Some patients will present not as chronic idiopathic constipation but as abdominal pain and be admitted for investigation.
‘Combination of treatment’ costs were calculated as the mean cost of the possible combination treatments available for each group. After successful manual disimpaction, all patients were assumed to be on a combination treatment for the rest of the 3 month initial treatment phase. Total costs of disimpaction and maintenance in the 3 month time frame were calculated for all possible pathways.
ii. Cost effectiveness of disimpaction by dose of a specific pharmacological treatment (polyethylene glycol plus electrolytes)
A decision analytic model was undertaken to model alternative PEG 3350 plus electrolytes doses in the treatment of disimpaction. Clinical outcomes and treatment doses came from a randomised controlled trial (RCT) conducted in the USA61 which aimed to investigate the effectiveness and safety of four different doses of PEG 3350 plus electrolytes in the treatment of childhood faecal disimpaction.
In the clinical trial, children were randomized into four groups and each group received a different daily dose (g/kg) of PEG 3350 plus electrolytes. shows the doses received by group and the proportion of children treated successfully (‘success rate’).
Doses and success rates, and cost per day.
For the economic model, these success rates were converted into QALYs. Estimates of the quality of life weights for constipation-related health states were obtained from economic literature reviewed earlier140 and are presented in . QALY values presented in this paper were used as they were elicited especially for the study from members of the general population using appropriate health economic methods (time trade off and standard gamble). The data were also from a large study of an adult population which was conducted in the UK.
Utility values for constipation-related health states.
The values used in the model developed for this guideline were 0.125 (3 months experiencing symptoms of constipation) and 0.235 (3 months well managed on medication). The QALY gain of moving from an unwell to a well-managed health state was 0.04 QALYs (0.225 to 0.185).
Cost data
To calculate the correct dose of treatment, we assumed a 25 kg child and calculated the corresponding number of sachets per day for the four groups. The cost of manual evacuation is reported in above. The cost of failure was modelled based on the following simplifying assumptions (see ):
Modelling pharmacological treatment for disimpaction: different doses.
Children who were still impacted after five days on dose 1 or 2 moved to dose 3. If this failed, it was assumed that a child underwent a successful manual evacuation.
Children who were still impacted after five days on dose 3 moved to dose 4. If this failed, they underwent a successful manual evacuation.
Children who were still impacted after five days with dose 4 repeated another five days of treatment with the same dose. If this failed, they underwent a successful manual evacuation.
Sachet doses and daily treatment costs were derived from BNFC (). Manual evacuation costs were derived from NHS reference costs 2007 ().
Cost effectiveness (incremental cost per disimpacted child and incremental cost per QALY) was undertaken from an NHS perspective. The time frame considered was the 5 days disimpaction period. The model applied is presented in .
iii. Pharmacological treatment for disimpaction: comparing different alternatives
The disimpaction model was developed assuming clinical equivalence of first line treatment for disimpaction to establish which group of pharmacological treatments, including all combinations of treatments and dose of treatments, including manual evacuation as a last resort for disimpaction, provided care at the lowest cost to the NHS over the initial 3 months of treatment. Using the clinical outcomes and resource used values obtained from GDG consensus, a model was constructed considering the decision to treat in primary care setting constipated children aged 2 to 11 years (to be consistent with the published economic63) with no flag to a serious underlying disorder after history and physical examination.
Different treatments pathways were proposed under four groups of pharmacological treatment strategies. For each treatment group, there were alternative decisions available if initial treatment with a baseline dose failed. The GDG specified all the different strategies (change of treatment, change of dose or combinations of treatments). For each group, the mean 3 month cost was calculated. This provided the GDG with information on which group of strategies provided the best value for money to the NHS given clinical equivalence.
The first treatment group (group 1) started with PEG 3350 plus electrolytes at a baseline dose. If the treatment was successful, the child stayed on this preparation, at half the dose, during the maintenance phase. If this baseline treatment failed, patients moved to a higher dose of PEG 3350 plus electrolytes. If the higher dose failed they then moved to a combination treatment with PEG 3350 plus electrolytes and sodium picosulfate. If all strategies from group 1 failed patients then moved to other treatment groups (2, 3 or 4). If all strategies from group 1 and another subsequent group failed the last choice treatment was manual evacuation.
The first choice treatment for group 2 was sodium picosulfate baseline dose. If this failed patients then moved to a higher dose. If the higher dose failed they then moved to one of three possible combinations: with PEG 3350 plus electrolytes; with senna; or with lactulose. If all strategies from group 2 failed the patients moved to other groups of treatment (1, 3 or 4). If all strategies from group 2 and another subsequent group failed the last choice treatment was manual evacuation.
For group 3 the first choice treatment was senna baseline dose. If this failed, patients then moved to a higher dose. If the higher dose failed they then moved to one of three possible combinations: with picosulfate; with lactulose; or with docusate. If all group 1 strategies failed the patients moved to other groups (1, 2, or 4). If all strategies from group 3 and another subsequent group failed the last choice treatment was manual evacuation.
For group 4 the first choice treatment was enemas. If it failed patients move to another group (1, 2 or 3). If group 4 and the subsequent group strategies failed the last choice treatment was manual evacuation.
In all, 136 possible treatment pathways were identified. A list of all 136 alternative pathways is presented at the end of the chapter.
Resource use
The analysis was undertaken from the NHS perspective and the time frame was the first three months of treatment. All pharmacological treatments were assumed to be administered at home, while a hospitalisation was required for enemas and manual evacuation procedures see above. Hospitalisations and GP/nurse outpatient visits following disimpaction were considered equal across the treatment options. Estimates of pharmacological treatment failure rates were agreed with the GDG members on a consensus base (). Daily doses and unit costs were derived from BNF children (last visited December 2008). When a range of doses was available, the lowest was considered as baseline dose. A higher dose was calculated applying a 25% increase to the baseline option, as advised by the GDG. Combinations of treatments included baseline doses for both options. Daily doses for the remaining maintenance period were calculated applying a 25% decrease to the disimpaction doses. Details of mean time to disimpaction, dosages, failure rate and hospitalisation unit costs are the same as those reported in . Total costs (for disimpaction phase, maintenance phase, and overall 3 month time frame) were calculated for all possible pathways and group options.
Mean times to disimpaction, failure rate, mean daily doses and hospitalisation unit costs.
Tree structure for the disimpaction model.
Resource use data, mean time to disimpact and failure rates for the different treatment options were obtained from discussions with the GDG ().
Effectiveness
In the first instance, the same level of clinical effectiveness for all first line treatments was assumed. For a specific combination of pharmacological treatments (docusate plus senna) offered when first line treatment had failed, clinical effectiveness was not assumed to be equivalent, but to be worse. Enemas also had a higher failure rate based on GDG consensus.
iv. Maintenance phase following disimpaction and initial management
An economic model for the maintenance phase of treatment post disimpaction was developed separately given the very large number of alternative pathways that would arise from combining the disimpaction and maintenance models. The model covered maintenance treatment (pharmacological and antegrade continent enema [ACE] procedure) for disimpacted children (age 2 to 11 years). The ACE strategy was included only as a last resort if other pharmacological strategies failed (see ). Each cycle covered a 3 month period after initial disimpaction. Results are reported after 3 months, at the end of 1 year (4 cycles) and 2 years (8 cycles). A discount rate of 3.5% was applied for the 2 year time frame*. ACE costs depend on which washout solution is used.
Maintenance model: reimpaction failure rates and costs applied to the maintenance model.
The pharmacological treatment strategies described in the disimpaction model were included (groups 1, 2 and 3) together with two additional treatments which are only offered in the maintenance phase: methylcellulose and liquid paraffin. This gave a total of 15 alternative strategies as first line treatment in the maintenance phase.
Drug doses were taken from BNFC (see ). All other healthcare resources and failure rates were agreed by GDG consensus. A decreased dose of 25% was applied to all successful disimpaction strategies to be continued as maintenance treatment. Three months disimpaction and maintenance costs are presented in . Compliance to treatment was also included in the model and adjustment to rate of success applied depending on whether the patients complied or not. For the purpose of this preliminary work a 100% compliance rate was considered for all treatments on offer.
Failure of one particular pharmacological strategy led to a switch to another alternative at the beginning of the following cycle. A maximum of eight different treatments were possible within the maximum length of 2 years. As instructed by the GDG, in the last 3 months of a completely successful maintenance period the doses were gradually decreased each month to 75%, 50% and 25%, respectively, before stopping.
The expected numbers of QALYs for the three time frames were estimated applying the same procedure as in the disimpaction model (see above).
The model is summarised in .
Tree structure for the maintenance model.
Results
i. Cost analysis by success rate for disimpaction
shows the range of costs associated with four alternative strategies for disimpaction of children with idiopathic constipation. These costs relate to the different starting doses published on the BNFC website (accessed December 2008). The table shows the drug costs and the total cost of care for the first 3 months of treatment starting with an initial baseline dose, moving to a higher dose if that fails, then a combination of pharmacological treatments, and finally manual evacuation as the last resort if all else fails. Once a treatment has been successful a maintenance dose of treatment is given for the rest of the 3 month period.
Results from costing hypothetical scenarios.
The results show that the treatment options using senna as the baseline drug resulted in lower overall costs compared with all other options. If effectiveness was the same for all treatments, this would be the least cost and therefore the most cost-effective option. However, if senna was not as effective as all the others, then all other treatments would be lower cost at medium or high levels of effectiveness, despite their higher drug prices. At these thresholds for effectiveness, there is no overlap in total costs between ‘success rate’ rows, indicating that if the GDG believes that one drug is effective at the medium (50%) or high (80%) level, then it will always be cheaper than one of the low-priced drugs at low level of effectiveness (20%). For all treatment options total costs were driven by success rate. High success implied a decrease in cost given the high cost of failure (that is, manual evacuation requiring hospitalisation).
The differences in effectiveness in the analysis were fairly large. The question therefore is how much more effective a higher cost drug (PEG+E) would have to be to offer a) cost effectiveness at the £20,000 per QALY threshold for cost effectiveness and b) cost saving.
Baseline scenario: we consider low dosages and low effectiveness rates (20%) for all treatments. PEG 3350 plus electrolytes would need to increase the effectiveness by 0.021 to be more cost effective than senna at the £20,000 per QALY threshold.
PEG 3350 plus electrolytes base dosage with 0.3 success rate is cheaper than senna base dosage with 0.2 success rate (£444 versus £501).
Cost effectiveness analysis of pharmacological treatment in the first 3 months of treatment, given £20,000 per QALY threshold.
suggests that the cost of a package of care does not alter greatly depending on the dose of treatment given. Total costs did not vary by more than 2% between the low dose and high dose preparations for any treatment, indicating that dose does not have a big impact on total cost. In fact, the cost of pharmacological treatment to treat impaction is dwarfed by the cost of failure when initial treatment fails. gives a graphic representation of this, showing that at all levels of success, the cost of success hardly registers on the chart next to the cost of failure. This is a strong indication that effectiveness is the dominant factor in determining the overall cost of treatment for disimpaction. Since success is determined by effectiveness and adherence to treatment, the treatment with the greatest chance of overall success should be the preferred option on cost-effectiveness grounds.
Cost saving threshold for pharmacological treatment in the first 3 months of treatment.
Cost of success and failure per treatment according to success rate (low dose only).
ii. Cost effectiveness of disimpaction by dose of a specific pharmacological treatment (PEG plus electrolytes)
The baseline cost analysis of PEG 3350 plus electrolytes by dose of treatment showed that dose 3 (1 g/kg, 4 sachets per day) was the preferred option. This is obvious since dose 3 costs less than the higher dose alternative (dose 4) but has the same reported level of effectiveness (see ).
Cost effectiveness analysis of treatment by dose of PEG 3350 plus electrolytes in the first 3 months of treatment, given £20,000 per QALY threshold.
Again, the data suggests a higher dose of treatment with higher success rate and higher short-term disimpaction costs (that is, cost of success, see dose 3) is more cost effective than lower doses at lower initial pharmacological costs which are less effective and therefore require costly intervention when they fail.
However, given the NICE threshold for cost effectiveness of £20,000 per QALY, the effectiveness of dose 4 has to rise by only 0.21% in order for this to be the preferred option, indicating that these results are highly sensitive to the effectiveness of the treatment (see ). illustrates that these results are driven by the cost of failure which is a higher proportion of the total costs than the drug costs themselves.
Decision modelling for disimpaction and initial maintenance: total costs over 3 months assuming equal effectiveness.
Total 3 month cost of success and failure, by dose of PEG 3350 plus electrolytes.
iii. Treatment for disimpaction: comparing different alternatives
Total costs per patient per group are reported in . Since effectiveness did not differ across pharmacological strategy groups, a cost minimisation exercise was considered. The treatment option with lowest costs was group 3 (senna, £73), followed by groups 2 (Picosulfate, £95) and 1 (PEG 3350 plus electrolytes, £97). The most expensive option was enemas (group 4, £1,208).
The cost results show that, using the treatment pathways suggested by the GDG, the difference in cost over 3 months between alternatives based on pharmacological treatments is around £20 to £25 per child. In this analysis, enemas are much less successful than pharmacological treatments (failure rates 75% and 20% respectively, see ) leading to higher use of manual evacuation as a last resort. The cost of enemas is high and is driven by the cost of 4 days hospitalisation which is based on GDG opinion of the likely treatment pathway for a child with idiopathic constipation.
Like the first cost model, senna is the cheapest treatment alternative based on its lower drug costs and assumed clinical equivalence. However, threshold analysis showed that the effectiveness of PEG 3350 plus electrolytes would have to be 2.6% higher than the next best alternative (in this case senna) in order for it to be the preferred option on cost-effectiveness grounds.
Cost effectiveness threshold analysis of disimpaction treatment and first maintenance in the first 3 months of treatment, given £20,000 per QALY threshold.
iv. Decision modelling for strategies for ongoing maintenance after disimpaction
Total costs and outcomes per patient per group are shown in . Since equal effectiveness across groups was assumed in the first instance, the differential costs of care only are reported with equal numbers of QALYs. The total cost for the first 3 months of maintenance treatment using PEG 3350 plus electrolytes at baseline dose is much higher than for any other pharmacological treatments at the baseline dose (over £70 where all other treatments are under £10). The only alternative that is equally as costly is a strategy of starting with picosulfate and switching to PEG 3350 plus electrolytes and sodium picosulfate where that fails.
Maintenance model: total costs and outcomes per patient after first 3 months of disimpaction and initial maintenance.
The cost per child of the treatment option using senna in the first cycle (3 months) is £2.70 and for PEG 3350 plus electrolytes it is £73. This is based on the cost of half the dose of treatment used in the first 3 months of disimpaction and initial maintenance, and is based on a strategy using more pharmacological options if a treatment fails before opting for a manual evacuation (requiring hospitalisation) as a last resort. In this model, fewer children require hospitalisation in the maintenance phase than in the disimpaction phase, reflected in lower costs overall for the same time period. This widened the gap between the cheapest option (senna) and the most expensive since the cost of hospitalisation was no longer the largest cost driver in the overall cost of treatment.
Using a modelling approach it was possible to calculate how much more effective a PEG 3350 plus electrolytes strategy would have to be in the maintenance phase (3 months, 1 year, 2 years) in order for it to be cost effective at the £20,000 per QALY threshold, and at what level of effectiveness a more expensive strategy would be cost saving. Since PEG 3350 plus electrolytes costs more in the maintenance phase, it needs to be more effective for it to be the preferred option. It has been reported earlier that higher priced therapeutic strategies with higher levels of effectiveness would become cheaper overall than strategies with lower initial drug costs. It is possible to estimate how much more effective PEG 3350 plus electrolytes would have to be in order for it to be preferred to all other strategies on cost-effectiveness grounds.
The analysis presented in suggests that an increase in effectiveness from 80% to just over 85% effectiveness in the first 3 months of treatment (and less in the longer term) would make PEG 3350 plus electrolytes the more favourable option.
Cost effectiveness threshold analysis of maintenance treatment, given £20,000 per QALY threshold after 3 months, 1 year and 2 years of treatment.
Conclusion
The effectiveness of pharmacological treatments to treat idiopathic constipation in children is not well established. The cost effectiveness of alternative pharmacological strategies (initial treatment with a baseline dose and alternative doses or combinations where that fails) can be modelled even where robust data is not available. The NICE threshold for cost effectiveness of £20,000 per QALY provides a decision rule that allows the GDG to consider how much more effective a more costly alternative would have to be in order for it to be preferred on cost-effectiveness grounds.
The results of the economic modelling can be summarised as follows:
The ‘cost of disimpaction by success rate’ model showed that treatments with a high chance (80%) of success cost less than treatment with a low chance of success (20%), regardless of the price of drugs used or the dose provided. Also, the cost of failure (changing doses, combining drugs and manual evacuation as a last resort) was a far greater determinant of overall cost than the cost of initial treatment.
The analysis by dose of PEG 3350 plus electrolytes showed that highly effective strategies will lead to cost savings due to the high downstream costs of invasive treatment requiring hospitalisation that are saved. Effectiveness is determined both by the type of drug used and by the dose given. The data we have been able to identify on doses of treatment suggest that higher doses of PEG 3350 plus electrolytes that lead to effectiveness levels of 95% compared with 55% for lower doses would be cost saving to the NHS.
The disimpaction model based on a consensus of treatment pathways developed by the GDG showed that oral pharmacological alternatives were more than ten times cheaper than enemas which were assumed to be less effective and require hospitalisation. At a 20% failure rate, oral pharmacological treatment provided a mean benefit of 0.23 QALYs per child. The threshold analysis showed that the effectiveness of PEG 3350 plus electrolytes would have to be 2.6% higher than the next best alternative in order for it to be the preferred option on cost-effectiveness grounds.
The maintenance model showed that, unlike the disimpaction model, the cost of drugs in the pharmacological treatment alternatives had a greater impact on the total of care than hospitalisation, which widened the gap between the cheapest and most expensive options.
The economic analysis used the clinical effectiveness evidence that was available, along with GDG opinion, to model the cost of the pharmacological treatment options available in the NHS to make the GDG's decisions more transparent. It is clear that treatment failure plays a major role in determining the total cost per child of disimpaction and maintenance so that the cheapest priced option is not the most cost effective overall. Not enough is known about the true difference in effectiveness between options, nor about how children's compliance with treatments that are effective when used properly impacts on the overall effectiveness of a particular treatment strategy. The economic analysis has shown that the treatment with the highest success rate is also likely to be the most cost-effective option, regardless of price.
