|
Optimising health care within given budgets: primary prevention of cardiovascular disease in different regions of Sweden |
Lofroth E, Lindholm L, Wilhelmsen L, Rosen M |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology The study compared three methods for preventing cardiovascular disease (CVD). The prevention strategies were:
smoking cessation through doctors' advice;
blood pressure-lowering drugs (only offered to individuals with a systolic blood pressure greater than 140 mmHg); and
lipid-lowering drugs (administered to individuals with a serum cholesterol level greater than 5.0 mmol/L).
It was reported that the drugs used to address high blood pressure problems were beta-blockers, diuretics, calcium antagonists, angiotensin-converting enzyme inhibitors and angiotensin. The lipid-lowering drugs used were not reported. Combinations of the three strategies were also evaluated.
Study population The study population comprised individuals aged between 40 and 69 years with increased risk of CVD, but free from CVD history, who were eligible for primary prevention with lipid-lowering, blood pressure-lowering drugs and advice on smoking cessation in three regions in Sweden (i.e. Goteborg, Vasterbotten and Ostergotland).
Setting Although this was a modelling study and the setting was not explicitly stated, it seems that the setting was the community and primary care. The economic study was carried out in Sweden.
Dates to which data relate The effectiveness data were derived from various studies published between 1990 and 2003. The cost and resource use data were derived from sources published between 1992 and 2003. All costs were reported for the price year 2000.
Source of effectiveness data The effectiveness data were derived from a review and synthesis of published studies.
Modelling The authors constructed a deterministic state-transition model to derive the cost-effectiveness ratios. Each individual free from CVD entered the model and was subject to an annual risk of myocardial infarction (MI), stroke or death from either, or other causes. The model was based on the assumption that once a person develops MI or stroke they cannot recover and will spend the rest of their life in the health states "MI" or "stroke". This assumption does not exclude the possibility that a person who has already experienced an MI cannot suffer a stroke and vice versa. It was also assumed that a person who had experienced an MI could die of stroke or any other cause. The time spent in the health states of stroke or MI was shaped by the 1-day, 1-year and 2-year probability of dying because of CVD. After a 2-year period, the probability of dying due to CVD was assumed to remain the same as the probability of dying in year two. The patients were followed up until their death or until they reached the age of 110 years.
According to computations of risk factors such as age, gender, smoking, serum cholesterol and systolic blood pressure, 108 mutually excluding groups were formed. Matching the risk group combination with all treatment options resulted in 864 different combinations. However, of these, at least one treatment combination was offered only to 342 combinations. The model was based on the assumption that the effects of prevention methods were multiplicative (i.e. the absolute risk reduction following two or three preventive methods equalled the product of each absolute risk reduction).
The authors also computed risk functions for MI and stroke. In the risk function, the incidence reflected the difference between the risk factor level in a studied group and the mean risk factor level in the population. The patients were assumed to receive drug prevention options for 5 years.
Outcomes assessed in the review The following parameters were used to calculate the risk functions:
time to first MI;
incidence of MI for a given sex at a given age;
cholesterol level;
mean cholesterol level for a given sex at a given age;
blood pressure level;
mean blood pressure level for a given sex at a given age;
smoker or non-smoker;
mean level of smokers for a given sex at a given age;
gender and age; and
the beta coefficient, which determines how much the risk differs from the average incidence for a certain age- and sex-specific group.
The following parameters were used in the model:
the mean reduction of cardiac events and ischaemic strokes due to lipid-lowering drug therapy,
the absolute risk reduction of MI and stroke due to blood pressure-lowering drugs, and
the reduction in the yearly risk of MI and stroke due to smoking cessation.
Study designs and other criteria for inclusion in the review The authors reported that clinical trials were included in the review. However, study designs and other inclusion criteria were not reported.
Sources searched to identify primary studies Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data Number of primary studies included Overall, the authors used approximately 16 primary studies as sources of effectiveness data.
Methods of combining primary studies The authors only derived mean estimates of effectiveness data from the available studies.
Investigation of differences between primary studies The authors do not appear to have investigated differences between the primary studies.
Results of the review The results of the review are too numerous to be reported here. All outcomes ascertained from the literature were fully reported.
Measure of benefits used in the economic analysis The authors used health utility (quality-adjusted life-years gained, QALYs) as a measure of benefit in the economic analysis. QALY weights were derived from the literature (Tengs et al. 2000, see 'Other Publications of Related Interest' below for bibliographic details).
Direct costs The health care costs included in the analysis were for anti-smoking counselling, consultation (including patients travel costs), blood pressure-lowering drugs, lipid-lowering drugs, and MI and stroke treatment during the first and second year. The costs were reported per patient but, in most cases, summary costs instead of unit costs were reported. Except for the case of anti-smoking counselling and consultation, the costs and the quantities were not reported separately. The costs and quantities of resources used were derived from published sources. Since the costs were incurred during more than 2 years, they were appropriately discounted at a rate of 3%. All costs were appropriately adjusted and reported for the price year 2000.
Statistical analysis of costs The costs were treated deterministically.
Indirect Costs The value of lost production was included in the analysis. It was computed as the difference between the annual gross income for patients with MI or stroke and the general population. The costs were derived from actual data and were appropriately adjusted and reported for the price year 2000. Since the costs were incurred during more than 2 years, they were appropriately discounted at a rate of 3%.
Currency Euros (EUR). Swedish kroner (SEK) were converted to Euros at a rate of EUR 1.00 = SEK 8.45.
Sensitivity analysis Sensitivity analyses were conducted, although it was unclear how many of the parameters were varied simultaneously. The tables suggest multi-way analyses investigating duration of treatment, discount rate and productivity losses, serum cholesterol levels, systolic blood pressure, age and productivity losses. The results were presented for smokers and non-smokers.
Estimated benefits used in the economic analysis The estimated benefits were reported in a league table for the population aged 40 to 60 years, which was stratified according to age, gender, smoking habits, serum cholesterol level, and whether or not patients received blood pressure-lowering drugs and cholesterol-lowering drugs. The estimated benefits for each stratum were too numerous to report here. The reader is referred to appendix A of the current study.
Cost results The total costs were reported in league table for the population aged 40 to 60 years, which was stratified according to age, gender smoking habits, serum cholesterol level, and whether or not patients received blood pressure-lowering drugs and cholesterol-lowering drugs.
The total costs for each stratum were too numerous to report here. The reader is referred to appendix A of the current study
Synthesis of costs and benefits The authors conducted a threshold analysis. Different thresholds were applied for each county. In Goteborg and in Ostergotland the threshold was EUR 1,687 per QALY gained, while in Vasterbotten it was EUR 6,192. Taking the optimal allocation of available resources into account, the authors reported that smoking cessation was the dominant method in all groups.
When production costs were excluded in the sensitivity analysis, of the 144 recommendations of the base-case analysis, 17 (12%) were altered. When the duration of treatment was varied, the threshold ratios varied from EUR 5,912/QALY (including cost of productivity losses, with treatment for 10 years and a discount rate of 3%) to EUR 17,549/QALY (excluding cost of productivity losses, with treatment for 1 year and a discount rate of 5%).
When comparing different league tables, the authors reported that the widest variation in cost-effectiveness ratios for hypertension treatment were observed in a 40-year-old non-smoking woman with a serum cholesterol level of 5.0 mmol/L and a systolic blood pressure of 150 mmHg. The cost-effectiveness ratios ranged from EUR 33,037 per QALY gained to EUR 129,217 per QALY.
The smallest variation in cost-effectiveness ratios was observed with the smoking cessation method offered to a 70-year-old woman with a serum cholesterol of 5.0 mmol/L and a systolic blood pressure of 139 mmHg. The cost-effectiveness ratios ranged from EUR 3,653 per QALY gained (including cost of productivity losses, at a discount rate of 3% and at any duration of treatment) to EUR 4,410 (excluding cost of productivity losses, at a discount rate of 5% and at any duration of treatment).
Authors' conclusions Misallocation of resources does take place. The authors suggested that resources allocated to smoking cessation should be increased by 124%, those allocated to hypercholesterolaemia treatment by 114%, and those allocated to blood pressure medical treatment should be reduced by 54%.
CRD COMMENTARY - Selection of comparators The authors restricted their analysis to smoking cessation, lipid-lowering drugs and blood pressure-lowering drugs. It was not clear why the comparators used were chosen, and no explicit justification was provided for the choice. You should decide if these represent valid comparators in your own setting.
Validity of estimate of measure of effectiveness The authors did not report whether a systematic review of the literature was undertaken. Although this might be common practice with models, it does not ensure that the best available data are used. The authors appear to have used data from the available studies selectively; the validity of the primary studies was not demonstrated and no weighting schemes were employed to reflect differences in sample sizes in different primary studies. Given these limitations, it is not possible to ascertain the validity of the estimates of effectiveness used in the model. Therefore, despite what would appear to be sound methodology for conducting the analysis, the uncertainty around how the model inputs were selected and used leaves the validity of the results in doubt.
Validity of estimate of measure of benefit The authors used health utility (QALYs) as a measure of benefit in the economic analysis. It is impossible to comment on the methods used to derive the health utility values as they were derived from a prior study. However, the utility values assigned to different health states were reported in the current study.
Validity of estimate of costs Although not explicitly stated, the inclusion of productivity losses would suggest that a wider perspective (i.e. societal) was adopted in the economic analysis. However, as the authors mainly reported summary costs, it is impossible to determine which aspects of the costs were included (e.g. overheads). In addition, the lack of unit cost estimates hinders the reproducibility of the results in other settings. The costs and quantities of resources were derived from published sources, but no statistical analysis of the estimates used was undertaken. The authors acknowledged that the reliability of the estimates used (e.g. drug costs) was uncertain, but did not address the issue further. All costs were appropriately adjusted and reported for the price year 2000, and discounting was correctly carried out.
Other issues The authors compared their findings with published data, reporting that their results were not in agreement with domestic and European guidelines: their results suggested that specific patient groups aged 40 years with a systolic blood pressure of 180 mmHg should be excluded from the intervention, while other groups (e.g. age 50 years with systolic blood pressure of 150 mmHg) should be included. However, the issue of generalisability of the results was not directly addressed. The authors do not appear to have presented their results selectively and their conclusions appear to reflect the scope of the analysis.
The authors reported a number of limitations to their study. First, the use of assumptions that were not well justified. Second, the conclusions reflected the amount of resources available in their setting, which should be taken into consideration in when generalising the results to other settings. Third, certain risk groups were small and, therefore, had insufficient statistical power to give robust results. Fourth, the proportion of drugs prescribed for primary and secondary prevention was based on rough estimates. Fifth, the study did not account for lifestyle-changing interventions, which might have proved to be superior to drug treatments, nor did it take alternative and possibly cheaper drugs treatments (e.g. diuretics) into consideration.
Implications of the study The authors suggested that, in the three regions studied, general reallocation of resources should take place by implementing a 60% reduction in blood pressure-lowering drugs and a redistribution of the resources to smoking cessation and to lipid-lowering drugs. The authors stressed the need for more comprehensive epidemiological information in order to derive robust estimates. In addition, the discussion highlighted several areas where further research is required.
Source of funding Supported by the Vardal Foundation and the Swedish National Institute of Public Health.
Bibliographic details Lofroth E, Lindholm L, Wilhelmsen L, Rosen M. Optimising health care within given budgets: primary prevention of cardiovascular disease in different regions of Sweden. 2006; 75: 214-229 Other publications of related interest Tengs TO, Wallace A. One thousand health-related quality-of-life estimates. Medical Care 2000;38:583-637.
Indexing Status Subject indexing assigned by NLM MeSH Adult; Aged; Budgets; Cardiovascular Diseases /prevention & Efficiency, Organizational /economics; Female; Health Care Rationing; Humans; Male; Middle Aged; Models, Statistical; Primary Health Care /economics /organization & Sweden; administration; control AccessionNumber 22006008018 Date bibliographic record published 31/08/2006 Date abstract record published 31/08/2006 |
|
|
|