CCBYNC Open access
Research

# Population based screening for chronic kidney disease: cost effectiveness study

BMJ 2010; 341 (Published 08 November 2010) Cite this as: BMJ 2010;341:c5869
1. Braden Manns, associate professor of medicine1234,
2. Brenda Hemmelgarn, associate professor of medicine 1234,
3. Marcello Tonelli, associate professor of medicine45,
4. Flora Au, research analyst24,
5. T Carter Chiasson, BHSc student4,
6. James Dong, research analyst45,
7. Scott Klarenbach, associate professor of medicine45
8. on behalf of the Alberta Kidney Disease Network
1. 1Department of Medicine, University of Calgary, Calgary, Alberta, Canada
2. 2Community Health Sciences, University of Calgary, Calgary
3. 3Libin Cardiovascular Institute of Alberta, University of Calgary, Calgary
4. 4Alberta Kidney Disease Network, Calgary
5. 5Department of Medicine, Division of Nephrology, University of Alberta, Edmonton, Alberta
1. Correspondence to: B Manns, Foothills Medical Centre, 1403 29th St. NW Calgary, Alberta Canada T2N 2T9 Braden.Manns{at}albertahealthservices.ca
• Accepted 2 September 2010

## Abstract

Objective To determine the cost effectiveness of one-off population based screening for chronic kidney disease based on estimated glomerular filtration rate.

Design Cost utility analysis of screening with estimated glomerular filtration rate alone compared with no screening (with allowance for incidental finding of cases of chronic kidney disease). Analyses were stratified by age, diabetes, and the presence or absence of proteinuria. Scenario and sensitivity analyses, including probabilistic sensitivity analysis, were performed. Costs were estimated in all adults and in subgroups defined by age, diabetes, and hypertension.

Setting Publicly funded Canadian healthcare system.

Participants Large population based laboratory cohort used to estimate mortality rates and incidence of end stage renal disease for patients with chronic kidney disease over a five year follow-up period. Patients had not previously undergone assessment of glomerular filtration rate.

Main outcome measures Lifetime costs, end stage renal disease, quality adjusted life years (QALYs) gained, and incremental cost per QALY gained.

View this table:

The cost of managing people with chronic kidney disease was assumed to be the incremental costs resulting from a new diagnosis, including those of nephrologist care, the requirement for angiotensin blockade, and laboratory monitoring (table 5). As admission to hospital in patients with stage 3-5 non-dialysis chronic kidney disease is usually because of comorbidity and because randomised trials of management of chronic kidney disease have not shown that interventions prevent admissions, we did not include the cost of admissions in our analysis. The annual cost of dialysis was estimated at $C64 218, assuming that nearly 81% of people were treated in haemodialysis centres.43 44 The cost of transplantation was estimated at$C84 531 for year one and $C35 545 for subsequent years.45 ### Valuing health benefits Health benefits were measured in cases of end stage renal disease prevented and QALYs gained. We did not assume any difference in utility estimates for “diagnosed” and “undiagnosed” chronic kidney disease, as studies of antihypertensive therapy (the only additional treatment in people with diagnosed chronic kidney disease) suggest no significant impact on quality of life.46 We estimated utilities for relevant health states (dialysis and transplantation) based on contemporary Canadian studies.45 47 ### Sensitivity and scenario analyses Sensitivity and scenario analyses were performed to determine the impact of uncertainty on the results of the model by varying all key parameters through plausible ranges. We conducted an additional scenario analysis examining the cost effectiveness of screening in people without diabetes but with known hypertension, reflecting an increased prevalence of chronic kidney disease in people with hypertension,10 as well as a doubling of the risk of progression to end stage renal disease among people with hypertension.48 As the cost of managing people with chronic kidney disease varies based on practice patterns, we also assessed the impact of using alternative cost estimates on results. To address limitations in classic univariate sensitivity analysis, we performed Monte Carlo simulation, which allows for the simultaneous sensitivity analysis of all variables over their plausible range.49 50 It does so by replacing estimates of probabilities, utilities, and costs with specific probability distributions, which are based on the reported means and variances for each variable. Statistical distributions were created around all of the variables for which there was substantial uncertainty of measurement, including use of a β distribution for proportions (that is, the risk of end stage renal disease and mortality), use of a normal distribution for normally distributed variables (that is, certain costs, relative risks, and utility measures), and log normal distribution for skewed variables (that is, certain costs). ## Results ### Baseline analyses Compared with no screening, population based screening for chronic kidney disease was associated with an incremental cost of$C463 and a gain of 0.0044 QALYs per patient overall, representing a cost per QALY gained of $C104 900 (table 6). In a cohort of 100 000 people, screening would be expected to reduce the number of people developing end stage renal disease over their lifetime from 675 to 657 (fig 1). In subgroups of people aged <65 and ≥65, the cost per QALY gained associated with screening was$C200 100 and $C93 700, respectively (table 6). Fig 1 Impact of screening for chronic kidney disease in cohort of 100 000 people on number of cases of end stage renal disease overall and for people with and without diabetes Table 6 Cost effectiveness of population based screening for chronic kidney disease and for targeted screening of high risk groups based on age, diabetes, and hypertension View this table: In subgroups of people with and without diabetes, the cost per QALY gained was$22 600 and $572 000, respectively. In a cohort of 100 000 people with diabetes, screening would be expected to reduce the number of people developing end stage renal disease over their lifetime from 1796 to 1741 (fig 1). In people without diabetes with and without hypertension, the cost per QALY gained was$C334 000 and $C1 411 100, respectively (table 6). ### Sensitivity analyses Tables 7, 8, and 9 show the results of sensitivity analyses conducted for the model overall, while tables 10 and 11 show the results of sensitivity analyses on the effectiveness of angiotensin blockade in patients with and without diabetes, respectively. Sensitivity analyses showed that screening restricted to people with diabetes was generally associated with a cost per QALY gained of <$25 000, which was robust to changes in all plausible variables (tables 7-10). Screening in people without diabetes, however, was associated with a cost per QALY of around $50 000 only under scenarios where identification of chronic kidney disease would result in use of treatments that could improve survival by at least 15%, and when the risk of progression to end stage renal disease in the absence of treatment was substantially higher than in the base case (table 11). Table 7 Sensitivity analysis of cost per QALY of screening for chronic kidney disease with varied rates of screening, use of angiotensin blockade, and adherence View this table: Table 8 Sensitivity analysis of cost per QALY for screening for chronic kidney disease (CKD) with varied incidence of disease and rates of progression to end stage renal disease (ESRD) in untreated patients View this table: Table 9 Sensitivity analysis of cost per QALY for screening for chronic kidney disease (CKD), exploring impact of variations in costs and quality of life View this table: Table 10 Sensitivity analysis of effectiveness of angiotensin blockade for management of CKD in people with diabetes View this table: Table 11 Sensitivity analysis of effectiveness of angiotensin blockade for management of CKD in people without diabetes View this table: ### Probabilistic sensitivity analysis Figures 2 and 3 show scatterplots of the incremental cost effectiveness of population based screening compared with no screening overall, highlighting uncertainty in the analysis, which is largely attributed to uncertainty regarding the impact of angiotensin blockade on mortality in people without diabetes and with chronic kidney disease. Given the current evidence of benefit for angiotensin blockade on both mortality and end stage renal disease in people with diabetes and chronic kidney disease, screening of people with diabetes led to better clinical outcomes in all simulations. Our analysis indicates a 37% probability that targeted screening for chronic kidney disease among people with diabetes is associated with a cost per QALY gained of <$20 000 and a 99% probability that the cost per QALY gained is <$50 000. Fig 2 Incremental cost effectiveness of population based screening for chronic kidney disease compared with no screening overall Fig 3 Incremental cost effectiveness scatterplot of population based screening for chronic kidney disease compared with no screening in people with diabetes ## Discussion Population based screening for chronic kidney disease is unlikely to be cost effective in unselected people or in those without diabetes. Although the prevalence of chronic kidney disease might be high enough to make screening worthy of consideration, particularly among older people,10 most people who would be found to have chronic kidney disease under a population based screening strategy would not have diabetes and would be likely to have non-proteinuric chronic kidney disease, relatively slow loss of kidney function, and low potential to benefit from angiotensin blockade.39 Therefore, the overall benefit of detecting and treating asymptomatic chronic kidney disease among people without diabetes is low. On the other hand, in people with diabetes, rates of progression to end stage renal disease are much higher, and there is strong evidence that angiotensin blockade reduces such progression 33 36 42 and improves survival.51 Given this, our study found that screening for chronic kidney disease in people with diabetes is associated with a cost per QALY in a range that is generally considered acceptable.52 53 54 ### Strengths and limitations We modelled progression of chronic kidney disease and mortality stratified by age, diabetes, and presence or absence of proteinuria using a large cohort of patients with chronic kidney disease followed over five years and estimated the benefit of screening from high quality meta-analyses of angiotensin blockade. Given that previous analyses examining the cost effectiveness of screening have used urine based screening methods,12 13 and as this might miss a considerable proportion of people with chronic kidney disease,10 we used estimated glomerular filtration rate to screen for the presence of chronic kidney disease. One limitation of our analysis is that we did not compare our results to a strategy of screening with urinalysis or a combined strategy of glomerular filtration rate and urinalysis, though all patients detected as having chronic kidney disease were assumed to undergo urinalysis. Proteinuria is a powerful predictor of an increased risk of end stage renal disease and death in people with chronic kidney disease.55 While screening with estimated glomerular filtration rate might identify more people than screening with urinalysis, most people identified are at lower risk of adverse outcomes and do not seem to gain additional benefit from angiotensin blockade.39 Our analysis was limited by the validity and availability of randomised trials (and meta-analyses) reporting the effectiveness of managing patients with chronic kidney disease. In general though, our analyses were guided by estimates of effectiveness from high quality meta-analyses. Screening in people without diabetes was unattractive given that angiotensin blockade does not seem to reduce mortality in people with non-diabetic renal disease.38 39 When we assumed that angiotensin blockade reduced mortality by 16% among people without diabetes (consistent with results of a large trial in which most participants did not have chronic kidney disease37), the cost per QALY for screening people without diabetes became more attractive ($40 800), though it is unclear whether this is a reasonable assumption. It could also be argued that people who are identified as having chronic kidney disease would be more likely to receive statins, which have been proved to improve survival in people with mild to moderate chronic kidney disease.56 57 58 59 In our analysis, however, we assumed that people at risk of cardiovascular events would already be receiving statins or aspirin (even without screening for chronic kidney disease). We therefore modelled only the impact of adding angiotensin blockade. In scenarios where we modelled the impact of treating identified patients with statins as well (resulting in a reduction in mortality of 16%58 at an additional cost of $620 a year), then the cost per QALY for screening in people without diabetes was$53 700.

We assumed that the incidence of end stage renal disease and death, which was determined in a cohort of people with known chronic kidney disease, would be similar to a cohort of previously unidentified patients with chronic kidney disease who were detected through screening. Although this assumption might overestimate, or underestimate, the true risks, our results were robust to plausible changes in these variables. We also assumed that people with chronic kidney disease identified by screening would not be receiving angiotensin blockade, which seems reasonable given that they had not previously had their glomerular filtration rate measured. Of note, the results did not vary significantly when up to 20% of patients screened were already receiving angiotensin blockade. Finally, our results are most relevant to the Canadian setting. Given that our prevalence estimates are based on a large North American cohort study,10 our estimates of efficacy based on randomised trials conducted throughout the world, and that Canada has publicly funded healthcare and relative costs that are likely similar to Europe, we think that our results are likely to be valid in other countries.

### Comparison with other studies

Previous studies have examined only the cost effectiveness of screening with urinalysis or urine based quantitative proteinuria.12 13 60 Like our study, these studies found that targeted screening of high risk groups (but not population based screening) might be cost effective.12 13 An analysis by Boulware et al suggested that screening for chronic kidney disease with measurement of proteinuria in patients with hypertension was associated with a cost per QALY of less than US \$20 000.12 These analyses, however, assumed that angiotensin blockade would improve survival in people without diabetes,12 an assumption not supported by available data in patients with chronic kidney disease.38 39 A study examining the cost effectiveness of population based screening for urinary albumin excretion, with treatment of those noted to have albumin excretion >15 mg/day with fosinopril to prevent cardiovascular events, based on the PREVEND study cohort, reported a cost per life year gained of €16 700.60

Population based screening for chronic kidney disease has been advocated by some,4 15 16 and the National Kidney Foundation/Kidney Dialysis Outcomes Quality Initiative clinical practice guidelines for chronic kidney disease have recommended screening for people at high risk of kidney disease, including those with diabetes and hypertension and aged over 60.1 14 Given that considerable resources would be required to set up a population based screening programme, our results are important. Our analyses suggest that screening for chronic kidney disease with estimated glomerular filtration rate is not cost effective in older people or in those without diabetes with hypertension but for people with diabetes is associated with a cost per QALY gained in the range of other funded interventions. We could not determine whether screening with urinalysis might be more cost effective overall or in certain subgroups compared with estimated glomerular filtration rate.

### Conclusions

Our results suggest that population based screening for chronic kidney disease with assessment of estimated glomerular filtration rate is not cost effective overall or in subgroups of people without diabetes but with hypertension or in elderly people. Targeted chronic kidney disease screening with estimated glomerular filtration rate in people with diabetes is associated with a cost per QALY that is similar to other publicly funded interventions.

#### What is already known on this topic

• Chronic kidney disease is common, and in many patients is asymptomatic and remains undiagnosed

• Previous analyses have examined the cost effectiveness of screening with urinalysis rather than estimated glomerular filtration rate, the standard method for diagnosing chronic kidney disease

• Population based screening for chronic kidney disease with estimated glomerular filtration rate assessment is not cost effective overall

• Targeted screening of people with diabetes is associated with an acceptable cost per QALY in publicly funded healthcare systems.

## Notes

Cite this as: BMJ 2010;341:c5869

## Footnotes

• Contributors: BM, SK, MT, and BH made substantial contributions to study conception and design, drafting, and critical revision. BM, SK, BH, FA, JD, and TCC were involved with data analysis and interpretation of data. BM is guarantor.

• Funding: This study was supported by an operating grant from Alberta Heritage Foundation for Medical Research (now Alberta Innovates-Health Solutions). The infrastructure required for this research was also supported by the Alberta Heritage Foundation for Medical Research Interdisciplinary Team Grants Program (which supports the Interdisciplinary Chronic Disease Collaboration). The funding organisations played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. BM, MT, and BH are supported by New Investigator Awards from the Canadian Institutes of Health Research and MT, SK, BM, and BH are supported by salary awards from Alberta Innovates-Health Solutions.

• Competing interests: All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: this project was supported by an operating grant from Alberta Heritage Foundation for Medical Research (now Alberta Innovates-Health Solutions) for the submitted work; no financial relationships with any companies that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.

• Ethical approval: This study was approved by the Calgary Health Region ethics board.

• Data sharing: No additional data available.

This is an open-access article distributed under the terms of the Creative Commons Attribution Non-commercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license. See: http://creativecommons.org/licenses/by-nc/2.0/ and http://creativecommons.org/licenses/by-nc/2.0/legalcode.

View Abstract