We therefore used an incidence approach [2] to retrospectively collect patient data. Our objectives were to document the extent and timing of long-term complications and recurrent thromboembolism and to estimate the health care costs of treating these conditions. Eight types of complications were defined (for controls, these complications are called events): superficial venous thrombosis, deep venous thrombosis, cellulitis, venous ulcer, varicose veins, stasis dermatitis, deep venous insufficiency, and pulmonary embolism.

Methods

Top Methods Results Discussion Author & Article Info References

The medical records of 257 patients with a history of deep venous thrombosis of a lower limb (verified by phlebography) and 241 controls without a history of thromboembolic disease were reviewed for clinical outcomes and use of health care resources. By using hospital diagnosis registries, we identified consecutive patients with deep venous thrombosis whose condition was diagnosed between 1970 and 1985 at Skovde County Hospital, Sweden, or 1 of its 18 associated outpatient settings. Patients who were registered in error, had thrombosis in a part of the body other than lower limb, or had thrombosis that had not been verified by phlebography were excluded. The 257 study patients were selected for thrombosis only. Controls were selected from among all persons living in Skovde, Sweden, on 31 December 1979; this date was chosen to allow a duration of follow-up similar to that for the patients with thrombosis. Controls were matched to the patients by age (the closest date of birth) and sex.

All patients who survived the follow-up period had been followed for at least 10 but no more than 15 years. Patients who died were followed until the year of death. Controls were followed for 15 years or until the year of death. At study entry, mean ages were 64 years in the thrombosis group (60% men) and 66 years in the control group (59% men).

The following complications or events were recorded: superficial venous thrombosis (verified by phlebography or clinical diagnosis); deep venous thrombosis (verified by phlebography); cellulitis, venous ulcer that was not caused by arterial insufficiency or diabetes, varicose veins, and stasis dermatitis (each verified by clinical diagnosis); deep venous insufficiency (verified by Doppler ultrasonography); and pulmonary embolism (verified by scintigraphy or autopsy). Total use of medical care for each recorded complication or event was ascertained from records of primary care and hospital care from six hospital departments. About 3000 medical records were surveyed to determine total number of outpatient visits, days in the hospital, diagnostic measures, surgical procedures, and pharmaceutical agents.

Costs of the health care resources were based on reported unit prices of inpatient and outpatient care and pharmaceutical agents. In Sweden, health care purchasers use pricing data to reimburse providers for cross-boundary health services. Thus, such billing data are available from local, regional, and national health services. From an economic point of view, pricing data cannot generally be considered a true measure of actual costs. However, true cost measures that reflect full-opportunity costs are not regularly produced in the health care system. In our study, we assumed that reported unit prices are reasonable approximations of costs. All estimated costs were calculated in Swedish kronors (SEK) in fixed 1990-1991 prices and then converted to U.S. dollars (in 1991, 1 U.S. dollar = SEK 5.5). A 5% discount rate was used to adjust for the different timings of complications or events.

To estimate the average present value of expected lifetime costs of treating post-thrombotic complications per patient of cohort i, discounted to the time of the primary thrombosis, we added all present and expected future costs according to the following formula: Equation 1 where TC is the total treatment cost for cohort i in year t after primary thrombosis, r is the discount rate, N is the sample population of cohort i, and i is 1 for patients with thrombosis and 2 for controls.

(1)

None of the funding parties had any role in the collection, analysis, or interpretation of study data.

Results

Top Methods Results Discussion Author & Article Info References

The results of our analysis are presented in Table 1 and Table 2. Table 1 shows the differences in the frequency of complications or events and the survival pattern for the two groups. At the end of the follow-up period, 242 post-thrombotic complications were recorded among the patients with thrombosis and 25 events were recorded among the controls (7 of these events were cases of primary deep venous thrombosis). In the patients with thrombosis, approximately two thirds of the complications occurred within 5 years after primary deep venous thrombosis had developed. The survival pattern also differed substantially between groups: Thirty-five percent of patients with thrombosis and 57% of controls were alive at the end of the period.

Because of differences in the frequency of complications or events and survival, the differences in the frequency of complications or events per patient exposed to risk were even larger. At the end of the follow-up period, a surviving patient in the thrombosis group had had, on average, 10 times more complications than a surviving control. The total costs of treating complications or events were SEK 7 850 696 ($1 427 399) for patients with thrombosis and SEK 607 104 ($110 383) for controls.

For the entire follow-up period, the average costs per complication or event were SEK 32 441 ($5898) for patients with thrombosis and SEK 24 284 ($4415) for controls. Venous ulcer was the most expensive type of complication, followed by deep venous thrombosis and pulmonary embolism (Table 2). For both groups, however, the treatment cost per complication or event varied greatly, indicating similar variation in the severity of the complications or events.

Among the patients with thrombosis, more than one third of the treatment cost was attributable to recurrent deep venous thrombosis. Among the controls, primary deep venous thrombosis accounted for 38.6% of the total treatment cost. The estimated average cost of primary deep venous thrombosis was SEK 33 455 ($6083) (Table 2).

At the end of the follow-up period, the discounted average present values of treatment cost were SEK 25 625 ($4659) for patients with thrombosis and SEK 2060 ($375) for controls. The difference in the discounted present values of treatment cost between the thrombosis and control groups (that is, SEK 25 625 –SEK 2060 = SEK 23 565 [$4285]) can be interpreted as the expected cost of treating post-thrombotic complications. Thus, an estimate of the present and expected cost of treating both primary deep venous thrombosis and related post-thrombotic complications is SEK 57 020 ($10 368) (that is, SEK 33 455 + SEK 23 565). According to this estimate, approximately 60% of the cost is attributable to primary deep venous thrombosis and 40% is attributable to post-thrombotic complications.

Discussion

Top Methods Results Discussion Author & Article Info References

In our study of long-term complications of deep venous thrombosis of the lower extremity, we compared patients who had phlebography-verified thrombosis with age- and sex-matched controls who served as a baseline benchmark. We considered the inclusion of controls to be important because the prevalence of nonthrombotic deep venous insufficiency increases with age [3]. One limitation of our study could be that we matched the groups for age and sex only; we were unable to match them for health conditions or factors that predispose patients to deep venous thrombosis or venous insufficiency. Nonetheless, even controls matched only for age and sex provide the incidence of deep venous insufficiency in an unselected population.

In our study, the post-thrombotic complications that occurred over 10 to 15 years of follow-up could be corrected for similar events among controls: In patients with thrombosis, the risk for a thrombotic complication was 10 times greater and the cost of such a complication was 12 times greater compared with controls. To our knowledge, ours is the first study to compare clinical and economic post-thrombotic complications with what could be considered the natural occurrence in an age- and sex-matched population.

In the control group, all patients were followed for 15 years or until death. In the thrombosis group, however, 30 patients (12%) were censored because they were followed for 10 to 14 years. These patients represent a loss of 3.7% of all years of follow-up. The number of complications per exposed patient decreased with time (Table 1). Through the 10 to 15 years of follow-up, the annual risk for a complication is about one third of the risk during the first 10 years. Thus, the number of complications in the thrombosis group is underestimated by about 1.2% (3.7% x one third) as a result of the censored years of follow-up. Because of the discounting principle, however, costs that are incurred after 10 years and are discounted by 5% will result in a present value of only about 50% of the original. Because of censored patients, therefore, we may have underestimated the present value of post-thrombotic complications by less than 1%. We did not adjust for these censored patients; thus, our estimate of the costs of post-thrombotic complications is conservative.

We obtained data on complications from patient records. This suggests that the frequency of complications was lower than that seen in Lindhagen and colleagues' study [3], in which each patient was investigated at follow-up. However, because our goal was to assess clinically important long-term consequences, our approach seems reasonable.

We estimated that the average cost per complication or event was approximately SEK 33 000 ($6000) for primary deep venous thrombosis and SEK 34 000 ($6182) for pulmonary embolism. In an earlier study of patients from Malmo General Hospital in southern Sweden [4], these estimates (converted to 1991 prices) were approximately SEK 20 000 ($3600) for deep venous thrombosis and SEK 24 000 ($4400) for pulmonary embolism. Because use of hospital resources and costs are local, our cost estimates (obtained from one hospital) may not be generalizable to other settings. One of our more general results is the ratio between the expected costs of long-term treatment of post-thrombotic complications and primary deep venous thrombosis. This ratio was approximately 0.75 (that is, the expected long-term costs of treating post-thrombotic complications are about 75% of the costs of treating primary deep venous thrombosis).

In the classic study by Bauer [5], the incidence of complications increased during the first several years of follow-up and decreased after about 5 years. However, decrease in incidence may start earlier when venous function is measured objectively [6]. During the first 5 years of our study period, complications developed in about two thirds of the patients with thrombosis and in 56% of the controls. Pulmonary embolism and recurrent thrombosis, however, usually occurred within the first year.

A high mortality rate in patients with thrombosis could be partially explained by concomitant cancer or cardiovascular disease [7, 8]. In our study, the mortality rate after 15 years was 65% in the thrombosis group and 43% in the control group.

Medical records contained little data on use of oral anticoagulation therapy, ulcer dressings, and supportive stockings. Thus, the costs of these measures are not included. Indirect costs resulting from loss of productivity were not estimated because no appropriate data were available. However, because two thirds of the patients were younger than 70 years of age at the time of thrombosis and most of them could be considered to be employed, the inclusion of indirect costs would further increase the long-term cost. This suggests that the estimated cost difference between the groups is minimal.

The estimated incidence of deep venous thrombosis in Nordic countries is 1.5 to 2 cases per 1000 persons per year; surgery without prophylaxis is an important risk factor [9, 10]. Untreated venous thromboembolism is associated with considerable risk for death and chronic venous insufficiency. Moreover, our findings show that long-term complications have a notable economic effect. Economic evaluations [4, 11, 12] have shown that prophylactic measures are cost-effective compared with no prophylaxis or surveillance and selective treatment of venous thromboembolism. We conclude that the use of measures to prevent thromboembolism and its long-term complications is justified on both clinical and economic grounds.

Mr. Jendteg: Institute of Health Economics, PO Box 2127, S-220 02 Lund, Sweden.

Dr. Johansen: Department of Surgery, Skovde Hospital, S-541 85 Skovde, Sweden.

Mr. Persson and Dr. Odegaard: Institute of Health Economics, PO Box 2127, S-220 02 Lund, Sweden.