Table 1.

Summary of findings for overall population: utility, RIO, or health state value information

Health state/outcome (categories of values and preferences)	Estimates of utilities	Certainty in evidence	Interpretation of findings
	No. of participants/studies
DVT* (Hogg et al,27,28 Lloyd et al,31 Locadia et al,32 Marvig et al,34 Utne et al37)	Range across studies: 0.61-0.99 Standard gamble: 0.81-0.99 Time trade-off: 0.84 VAS: 0.65-0.72 EQ-5D utility: 0.61-0.79 SF-6D: 0.64	⊕⊕⊕○ Moderate certainty due to inconsistency†	People may probably find DVT having a moderate or a trivial impact on their lives. There is likely an important variability for this assessment.
	1702 participants from 6 studies‡ Standard gamble: 260 participants from 2 studies Time trade-off: 124 participants from 1 study VAS: 485 participants from 3 studies EQ-5D utility: 1318 participants from 4 studies SF-6D: 44 participants from 1 study
PE* (Hogg et al,27,28 Lloyd et al,31 Locadia et al,32 Marvig et al,34 Tavoly et al,36 Utne et al37)	Range across studies: 0.63-0.93 Standard gamble: 0.75-0.93 Time trade-off: 0.63 VAS: 0.67-0.70 Rating scale modeled: EQ-5D: 0.621-0.80 SF-6D: 0.68	⊕⊕⊕○ Moderate certainty due to inconsistency†	People may probably find PE having a moderate or a small impact on their lives. There is likely an important variability for this assessment.
	1474 participants from 5 studies§ Standard gamble: 260 participants from 2 studies Time trade-off: 124 participants from 1 study Rating scale modeled: EQ-5D: 877 participants from 1 study VAS: 257 participants from 2 studies EQ-5D: 213 participants from 1 study SF-6D: 44 participants from 1 study
PTS (Locadia et al32)	0.82 (IQR, 0.66-0.94)	⊕⊕⊕○ Moderate certainty due to indirectness¶	People may probably find PTS having a small impact on their lives. There is likely an important variability for this assessment.
	124 participants from 1 study based on time trade-off
Mild PTS (Lenert and Soetikno,30 O’Meara et al35)	Range across studies: 0.99-1.00	⊕⊕○○ Low certainty due to indirectness and imprecision||,#	People might possibly find mild PTS having a trivial impact on their lives. There is likely no important variability for this assessment.
	66 participants from 2 studies based on standard gamble
Severe PTS (Lenert and Soetikno,30 O’Meara et al35)	Range across studies: 0.93-0.98	⊕⊕○○ Low certainty due to indirectness and imprecision||,#	People might possibly find severe PTS having a trivial or a small impact on their lives. There is likely no important variability for this assessment.
	66 participants from 2 studies based on standard gamble
Gastrointestinal tract bleeding event (Hogg et al,27 Lloyd et al,31 Locadia et al32)	Range across studies: 0.59-0.65 Standard gamble: 0.65 (IQR, 0.15-0.86) Time trade-off: 0.65 (IQR, 0.49-0.86) Rating scale modeled: EQ-5D 0.59 (95% CI, 0.46-0.69)	⊕⊕⊕○ Moderate certainty due¶ to indirectness**	People may probably find gastrointestinal tract bleeding having a moderate impact on their lives. There is likely an important variability for this assessment.
	1217 participants Standard gamble: 216 participants from 1 study Time trade-off: 124 participants from 1 study Rating scale modeled: EQ-5D: 877 patients from 1 study
Muscular bleeding event (Locadia et al32)	0.76 (IQR, 0.59-0.95)	⊕⊕⊕○ Moderate certainty due to indirectness¶	People may probably find a muscular bleeding event having a moderate impact on their lives. There is likely an important variability for this assessment.
	124 participants from 1 study based on time trade-off
Central nervous system bleeding (including intracranial bleeding) (Lenert and Soetikno,30 O’Meara et al35)	Range across studies: 0.29-0.60	⊕○○ Very low certainty due to inconsistency, indirectness, and imprecision||,#,††	People appear to find central nervous system bleeding having an important or a moderate impact on their lives. There is likely an important variability for this assessment.
	66 participants from 2 studies based on standard gamble
Major intracranial bleeding event (Hogg et al27)	Standard gamble: 0.15 (IQR, 0.00-0.65)	⊕⊕⊕⊕ High certainty	People find major intracranial bleeding having a large impact on their lives. There is likely an important variability for this assessment.
	216 participants from 1 study based on standard gamble
Minor intracranial bleeding event (Hogg et al27)	Standard gamble: 0.75 (0.55-0.92)	⊕⊕⊕⊕ High certainty	People find minor intracranial bleeding having a moderate impact on their lives. There is likely an important variability for this assessment.
	216 participants from 1 study based on standard gamble
Treatment with VKAs (Dranitsaris et al,25 Keita et al,29 Locadia et al,32 Marchetti et al33)	Time trade-off: range 0.38 to 0.99, EQ-5D VAS: mean 0.61	⊕⊕⊕○ Moderate certainty due to inconsistency and indirectness‡‡	People may probably find treatment with VKAs having an important and a trivial impact on their lives. There is likely an important variability for this assessment.
	296 participants in 4 studies (3 for time trade-off, 1 in EQ-5D VAS)
Treatment with LMWH (Dranitsaris et al,25 Marchetti et al33)	Range: 0.66-0.99	⊕⊕○○ Low certainty due to inconsistency/indirectness and imprecision‡‡,a	People might possibly find treatment with LMWH having a moderate or trivial impact on their lives. There is likely an important variability for this assessment.
	72 participants in 2 studies based on time trade-off
Treatment with standard systemic thrombolysis (Enden et al26)	0.84 (95% CI, 0.81-0.87)	⊕⊕⊕○ Moderate certainty due to imprecisiona	People may probably find treatment with thrombolysis having a small impact on their lives. After 24 mo, there were probably no differences in treatment burden between the standard treatment arms and the additional CDT. There is likely no important variability for this assessment.
	99 participants from standard treatment group (n = 189), 1 study based on EQ-5D utility
Treatment with CDT (Enden et al26)	0.80 (95% CI, 0.75-0.85)	⊕⊕⊕○ Moderate certainty due to imprecisiona
	90 participants from CDT treatment group (n = 189), 1 study based on EQ-5D utility

High certainty in evidence: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

CDT, catheter-directed thrombolysis; CI, confidence interval; IQR, interquartile range, VAS, visual analog scale.

^*Utilities measured with VAS, time trade-off, and standard gamble, as well as EQ-5D and SF-6D utility.

^†Of the 6 studies included, we can observe heterogeneity around the estimates across different studies. Study participants or measurement methodology alone could not explain the differences.

^‡Hogg et al²⁸ reported standard gamble, VAS, and SF-6D utility information; Marvig et al³⁴ and Utne et al³⁷ reported EQ-5D VAS and EQ-5D utility; and Lloyd et al³¹ reported VAS and EQ-5D utility.

^§Hogg et al²⁸ reported standard gamble, VAS, and SF-6D utility information; Lloyd et al³¹ reported VAS and EQ-5D utility.

^¶Locadia et al³² is a cross-sectional study interviewing participants with decision analysis. This study included 3 groups of patients, which may compromise the applicability of the overall results: (1) newly diagnosed patients with a first or second episode of VTE for whom treatment with VKAs had been started, (2) patients who had experienced an episode of major bleeding during treatment with VKAs in the previous year, and (3) patients with a PTS, diagnosed ≥1 year after an episode of DVT, who had been treated with VKAs for ≥3 months. The latter 2 patient groups were patients who have had the experience and, therefore, were not the optimal population who would be at risk for decision making.

^||The included studies have different populations than the patients facing the choice: 30 community volunteers (Lenert and Soetikno³⁰), 36 patients with DVT (16/36 participants) and without DVT (O'Meara et al³⁵).

^#In total, the sample size was too small; there were only 66 participants (excluding 30 physicians) from 2 studies.

^**Downgraded due to indirectness. Locadia et al³² included different group of population of patients who had experienced VTE, major bleeding and post–thrombotic syndrome, and Lloyd et al³¹ reported on major bleeding, rather than specific gastrointestinal bleeding.

^††The estimates from the 2 studies were very different, although both studies used standard gamble and measured the same outcome.

^‡‡Downgraded considering both quality criteria, because the observed variability (people finding taking VKAs as having both important or small impact on their lives) might be explained by the variability in the included populations across studies (people from the general population [Dranitsaris et al²⁵]); patients attending the local anticoagulation clinic (Marchetti et al³³); and patients with newly diagnosed first or second episode of VTE, patients with an episode of major bleeding during treatment the previous year, and patients with a PTS after an episode of DVT treated for ≥3 months (Locadia et al³²), as well as VTE outpatients receiving anticoagulation therapy by VKA or DOAC (Keita et al²⁹).

^aSmall sample size of included study or subgroup of patients.