|Computer simulated cost effectiveness of care management strategies on reduction of long-term sequelae in patients with non-insulin dependent diabetes mellitus
|Demers D, Clark N, Tolzmann G, MacLean C, Benedini K, Farnham P, Plant-DeHayes A, Nagy P
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.
A series six care management strategies for the care of patients with non-insulin dependent, or type II, diabetes mellitus (NIDDM). The strategies were designed to serve NIDDM people based on the vision of patient self-care. The six proposed initiatives were as follows: (1) all patients with diabetes should receive comprehensive education in diabetes self-management at onset and be afforded the opportunity for regular, ongoing education over their lifetime; (2) each patient should have a care passport that prompts regular care activities, provides a patient-held mini-record of health status for patients and provides coupons for discounts if treatment dates are met; (3) all patients should have access to adequate medical supplies including glucose meters, test strips, insulin, and other items provided based on ability to pay; (4) a sick day protocol and diabetes help line with 24-hour care guidance for unexpected episodes should be available to patients when they become hyper or hypoglycemic; home health care professionals, linked to the help line, should be on call 24 hours a day to support at home care of sick day episodes; (5) regular peer support groups should be performed so that patients can encourage and support each other; (6) definitive care standards should be established for both the outpatient and inpatient setting to eliminate unwarranted variation in practice patterns and to ensure maximum value patient care. The patients' interest in, and support for these initiatives was established through multiple patient support groups.
Type of intervention
Secondary prevention and treatment.
Economic study type
Hypothetical patients with non-insulin dependent, or type II, diabetes mellitus (NIDDM). The target population for the study was reported to be a self-contained, functional community (a population of about 136,000) who were predominantly white (98%) and which was rural in nature.
Community, primary, and secondary care. The economic study was carried out in the USA.
Dates to which data relate
The dates during which the effectiveness and resource use data were compiled were not explicitly specified. The model covered a 15-year period from 1996 to 2011. The price years do not appear to have been reported.
Source of effectiveness data
The evidence for the final outcomes consisted of assumptions made by the authors based on information gathered from hospitals, physicians, offices, home health care agencies, the State Health Department, the state affiliate of the American Diabetes Association, and other sources.
A computer model was developed to simulate the effect of the adopted care strategies on glycemic control, the impact of control on complications, and the reductions in morbidity and cost over a 15-year period. The simulation software used was 'I think (R)'.
Methods used to derive estimates of effectiveness
Assumptions made by the author based on information gathered from hospitals, physicians, offices, home health care agencies, the State Health Department, the state affiliate of the American Diabetes Association, and other sources.
Estimates of effectiveness and key assumptions
The simulation model variables and values were as follows:
population growth rate per year in the study area, 0.01;
diabetes mellitus (DM) prevalence rate, 0.026;
initial stock of DM patients, 3,555;
DM incidence rate, 0.0022;
mortality rate in DM population: aged 45-54 years, 0.000113, aged 55-64 years, 0.000327, aged 65-74 years, 0.000737, aged 75-84 years, 0.001439, aged over 85 years, 0.002526;
DM population with education at onset, 0.10;
base incidence rate of retinopathy, 0.00049;
DM retinopathy rate multiple, 10.3;
base incidence cardiovascular (CV) disease, 0.0025;
base incidence DM CV disease rate multiple, 5.1;
base incidence end-stage renal disease (ESRD), 0.0000138;
DM ESRD rate multiple, 3.3;
base incidence neuropathy, 0.00023;
DM neuropathy rate multiple, 3.3;
impact of control on complications, 1:0.16;
effect of tenure, .1:1.7;
education effect/total effect, 30%;
care standards/total effect, 30%;
supplies effect/total effect, 25%;
sick day protocol effect/total effect, 5%;
peer support effect/total effect, 10%.
It was estimated that 40% of the NIDDM population was undiagnosed. The proportion of the NIDDM population in acceptable control of their blood sugars (HgbA1c less than 8%) was assumed to be equivalent to 25%.
Measure of benefits used in the economic analysis
The benefit measures were the total number of patients with NIDDM in the study area, the percentage of the NIDDM population in acceptable glycemic control, the number of patients with complications. The calculations were based on assuming a programme coverage rate of 20% in year 1, 50% in year 2, 75% in year 3, 80% in year 4, and 80% in years 5-15 (due to patient migration and non-compliance factors).
Costs were not discounted despite a time frame of 15-years in which the costs occurred. Quantities appear not to have been reported separately from the costs. Cost items were reported separately. The cost analysis covered the costs of complications plus the start-up cost and ongoing cost of the care management initiatives proposed. The perspective adopted in the cost analysis appears not to have been explicitly reported. The information incorporated in the model was gathered from different sources. The price year was not explicitly reported.
It appears that a series of one-way sensitivity analyses ('what-if' analysis as mentioned in the paper) was conducted on the percentage of the NIDDM population using one or more care management initiatives, the assumed impact each care management initiative would have on control, the cost of providing each care management strategy to the population served, the graduated effect of increasing control on complications, and the cost per complication by types.
Estimated benefits used in the economic analysis
In the base case scenario (no case management strategies), the total number of patients with NIDDM in the study area grew from 3,555 in 1996 to 3,627 in 2011, remaining relatively constant at about 2.3% of the total population.
In the care management scenario, the size of the population in the year 2011 was 3,634 (due to seven fewer deaths within the population because of the effect of glycemic control on overall mortality).
In the care management scenario, the percentage of the NIDDM population in acceptable glycemic control grew from 30% in year 1 to 80% in year 10 (and was constant at 80% thereafter).
In the base case scenario, the number of patients with complications rose from 17% in year 1 to 20% in year 15, reflecting the effect of disease tenure.
In the care management scenario, the number of patients with complication fell to 5% in year 15, reflecting the impact of better glycemic control.
Over the 15-year time horizon of the model, the total number of patients with complication was 5,732 less in the case management scenario compared to the base case.
The NIDDM care management initiatives, when fully operational in year 3 of the study, would require a total annual investment of $641,000. Costs beyond year 3 grew relatively due to the modest increase in the NIDDM prevalence over the study period. The total care management costs over the 15-year period equalled $9.5 million.
In the base case scenario, total cost of complications grew from $10.7 million in year 1 to $14.9 million in year 15. In the care management scenario, the corresponding value fell to $3.8 million. Over the 15-year period, this represented a $121.9 million difference.
The net cost savings amounted to $112.4 million over the study period, an average return on investment (ROI) of $7.5 million per year. A positive ROI was not observed until year 3 and it was maximised in years 6 and grew thereafter relative to the number of persons with NIDDM enrolled in the care management initiatives.
Synthesis of costs and benefits
Costs and benefits were not combined since the care management scenario was the dominant strategy over the study period.
Improvements in the control of diabetes can be seen as an investment that hopefully results in fewer complications, better health, and reduced costs. This study suggests that the greatest benefits can be expected from improvements in the management of the disease at its earliest stages.
CRD COMMENTARY - Selection of comparators
The policy of using no care management strategies was regarded as the comparator. It allowed the active value of the care management strategies to be evaluated.
Validity of estimate of measure of effectiveness benefit:
The internal validity of the effectiveness results cannot be objectively assessed as insufficient information was provided regarding the sources of the effectiveness data used in the simulation model. However, performing sensitivity analysis on some effectiveness parameters may have enhanced its validity.
Validity of estimate of measure of benefit
The analysis of benefits was modelled. The instrument used to derive a measure of health benefit, a computer simulation model, was appropriate.
Validity of estimate of costs
Some of the positive features of the cost study contributing to its validity were as follows: adequate details of the methods used in the cost estimation were given; sensitivity analyses were performed on some of the cost parameters. However, the following limitations may have adversely affected the validity: quantities were not reported separately from the costs; it is not entirely clear whether costs data were true costs or charge data; the price year and perspectives adopted in the cost analysis were not reported; the effects of indirect costs (lost opportunity) were not included.
The authors' conclusion should be evaluated in the light of the inherent limitations of modelling exercises. The issue of generalisability to other settings or countries was not explicitly addressed. Some comparisons were made with other studies. No discounting was applied to costs or benefits despite the 15-year time frame of the analysis. Regarding the chronic nature of DM disease, a cost-utility approach, would have produced more meaningful information by incorporating the subjective assessment of the patients in the analysis.
Implications of the study
It has been said that care management for patients with NIDDM represents the "canary" of care management programmes. If the health care industry is successful in demonstrating positive results with these patients, other groups will follow.
Demers D, Clark N, Tolzmann G, MacLean C, Benedini K, Farnham P, Plant-DeHayes A, Nagy P. Computer simulated cost effectiveness of care management strategies on reduction of long-term sequelae in patients with non-insulin dependent diabetes mellitus. Quality Management in Health Care 1997; 6(1): 1-13
Subject indexing assigned by NLM
Blood Glucose /analysis; Computer Simulation; Cost Control /methods; Cost-Benefit Analysis; Delivery of Health Care, Integrated /organization & Diabetes Mellitus, Type 2 /complications /economics /epidemiology /therapy; Disease Management; Humans; Models, Econometric; Outcome Assessment (Health Care); Patient Care Planning; Vermont /epidemiology; administration
Date bibliographic record published
Date abstract record published