• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Maund E, Craig D, Suekarran S, et al. Management of Frozen Shoulder: A Systematic Review and Cost-Effectiveness Analysis. Southampton (UK): NIHR Evaluation, Trials and Studies Coordinating Centre (UK); 2012 Mar. (Health Technology Assessment, No. 16.11.)

Appendix 14Exploratory mapping analysis

Mapping from SF-36 PCS and MCS onto EQ-5D

Pattern of EQ-5D scores and SF-36 (PCS, MCS) scores within the SAPPHIRE data set (for complete measurements at 1 month)

EQ-5D groupingnEQ-5D (mean)SF-36 PCS (mean)SF-36 MCS (mean)
< 011−0.0728.1127.46
0–0.249230.1133.4740.18
0.25–0.49940.2939.8652.83
0.5–0.699600.6236.9643.82
0.7–0.799350.7643.7853.18
0.8–0.899200.8150.0851.43
0.9–1.00NANANA
Full index18152.4855.05
Overall1710.5940.4946.19

NA, not applicable.

Pattern of EQ-5D scores and SF-36 (PCS, MCS) scores within the SAPPHIRE data set (for complete measurements at 3 months)

EQ-5D groupingnEQ-5D (mean)SF-36 PCS (mean)SF-36 MCS (mean)
< 010−0.0735.9645.70
0–0.249180.1139.1847.35
0.25–0.49910.2636.6441.14
0.5–0.699440.6339.6842.01
0.7–0.799210.7544.0750.23
0.8–0.899280.8242.9153.2
0.9–1.00NANANA
Full index111.040.6848.95
Overall1330.5940.7947.21

NA, not applicable.

Pattern of EQ-5D scores and SF-36 (PCS, MCS) scores within the SAPPHIRE data set (for complete measurements at 12 months)

EQ-5D groupingnEQ-5D (mean)SF-36 PCS (mean)SF-36 MCS (mean)
< 011−0.1227.8129.83
0–0.249110.0932.0038.27
0.25–0.49910.2621.5147.41
0.5–0.699490.6134.5243.54
0.7–0.799300.7545.5448.71
0.8–0.899320.8250.1353.35
0.9–1.00NANANA
Full index29154.8755.79
Overall1630.6642.5347.34

NA, not applicable.

Pattern of EQ-5D scores and SF-36 PCS and MCS scores within the SAPPHIRE data set (for complete responses) at different time points

FIGURE 30. One-month follow-up.

FIGURE 30One-month follow-up

(a) EQ-5D and SF-36 PCS, (b) EQ-5D and SF-36 MCS

FIGURE 31. Three-month follow-up.

FIGURE 31Three-month follow-up

(a) EQ-5D and SF-36 PCS, (b) EQ-5D and SF-36 MCS.

FIGURE 32. Twelve-month follow-up.

FIGURE 32Twelve-month follow-up

(a) EQ-5D and SF-36 PCS, (b) EQ-5D and SF-36 MCS.

Regression models using main effects with and without squared terms and interaction term (using individual-level data at 1, 3 and 12 months)

Model 1Model 2Model 3Model 4Model 5
Econometric estimation methodOLSaOLSaOLSaTOBITCLAD
Independent variablesPCS MCSPCS MCS, squared termsPCS MCS, squared terms and interaction termPCS MCS, squared terms and interaction termPCS MCS, squared terms and interaction term
Coeff.SECoeff.SECoeff.SECoeff.SECoeff.SE
Intercept−0.26450510.004122−0.9479650.1748065−1.3784580.0449899−1.1711230.3070996−1.8397970.4490999
PCS0.01397560.0015533−0.02012760.01136770.03357920.00808140.0232850.01082020.05680590.0155343
MCS0.00655440.22873040.03660590.00817610.044060.01179330.04158460.00954590.04465960.0092294
PCS*PCS−0.0000910.000089−0.00010660.0000965−0.00001060.0001109−0.0003320.0001687
MCS*MCS−0.0034440.0001092−0.0003220.0001171−0.00032380.0001014−0.00026610.0000738
PCS*MCS−0.00025280.0000856−0.00016930.0001162−0.00036470.0001247
Adjusted R20.38400.41620.4284
ME−0.0025−0.0039−0.0020−0.0235−0.0265
MAE0.18890.18710.18610.18770.1815
RMSE0.2548990.2591120.259060.265080.266589
|Differencel|b
|Δ| ≤ 0.1015482%15181%15482%15683%15482%
|Δ| ≤ 0.0513673%13070%13770%13974%14578%
|Δ| ≤ 0.018545%7641%8847%9149%10255%
ActualPredictPredictPredictPredictPredict
Mean0.62190.62440.62580.62390.64540.6484
SD0.291910.196330.206830.209390.229370.21983
Min.−0.180.20−0.10−0.09−0.10−0.07
Max.1.000.920.900.921.010.86
Rangec1.180.721.011.011.110.93

SE, robust standard error.

a

With adjustment for repeated measurements at 1, 3 and 12 months' follow-up.

b

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

c

Range = maximum value - minimum value.

Total n = 467. Estimation data set = 280 (60%), validation data set = 187.

Regression models using the main effects with and without squared terms and interaction term (using models estimated from 3-month data and used to predict EQ-5D scores at 12 months)

Model 1Model 2Model 3Model 4Model 5
Econometric estimation methodOLSaOLSaOLSaTOBITCLAD
Independent variablesPCS MCSPCS MCS, squared termsPCS MCS, squared terms and interaction termPCS MCS, squared terms and interaction termPCS MCS, squared terms and interaction term
Coeff.SECoeff.SECoeff.SECoeff.SECoeff.SE
Intercept0.28426820.1811317−0.62335330.5770837−1.5524340.749881−1.8148380.7887539−0.67704891.035821
PCS0.0045590.0034250.00180810.02017210.02497010.02329690.02865080.02445270.01572590.0273862
MCS0.00263230.00307360.05162040.02219070.0734180.02466420.08316820.02599280.03996890.0316185
PCS*PCS0.00001860.0002580.00004640.00025410.0000410.00026830.00001520.0002815
MCS*MCS−0.00055610.0002495−0.00058770.0002459−0.00067110.0002584−0.00028750.0002772
PCS*MCS−0.000520.0002744−0.00058160.0002883−0.00026410.0003168
Adjusted R20.01470.05130.0830
ME0.06620.07030.07610.0654−0.0100
MAE0.19690.20450.20390.26720.1531
RMSE0.2479360.2668840.2600140.3511590.221678
|Differencel|b
|Δ| ≤ 0.105131%5534%5836%5836%9458%
|Δ| ≤ 0.052314%3421%3924%3622%5634%
|Δ| ≤ 0.0174%159%88%176%1610%
ActualPredictPredictPredictPredictPredict
Mean0.66900.60280.59870.59290.60370.6790
SD0.292540.072500.12770.158070.196920.12368
Min.−0.240.46−0.20−0.24−0.500.10
Max.1.000.710.760.860.910.84
Rangec1.020.2560.6070.7550.8460.541

SE, robust standard error.

a

With adjustment for repeated measurements at 1, 3 and 12 months' follow-up.

b

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

c

Range = maximum value - minimum value.

Validation data set = 163.

Actual versus predicted EQ-5D: mapping SF-36 PCS and MCS to EQ-5D

FIGURE 33. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS1 model.

FIGURE 33Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS1 model

FIGURE 34. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS2 model.

FIGURE 34Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS2 model

FIGURE 35. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS3 model.

FIGURE 35Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS3 model

FIGURE 36. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, TOBIT model.

FIGURE 36Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, TOBIT model

FIGURE 37. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, CLAD model.

FIGURE 37Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, CLAD model

Mapping from pain visual analogue scale onto EQ-5D

Pattern of EQ-5D scores and pain visual analogue scale scores within the SAPPHIRE data set (for complete measurements at 1 month)

EQ-5D groupingnEQ-5D (mean)Pain VAS (mean)
< 011−0.0780.91
0–0.249230.1158.26
0.25–0.49940.2926.25
0.5–0.699680.6351.00
0.7–0.799340.7633.91
0.8–0.899200.8129.4
0.9–1.0191.019.74
Full index191.019.74
Overall1790.5944.13

Pattern of EQ-5D scores and pain visual analogue scale scores within the SAPPHIRE data set (for complete measurements at 3 months)

EQ-5D groupingnEQ-5D (mean)Pain VAS (mean)
< 08−0.0518.13
0–0.249150.129.27
0.25–0.49910.2650
0.5–0.699520.6341.87
0.7–0.799240.7540.17
0.8–0.899260.8338.88
0.9–1.0151.034.67
Full index151.034.67
Overall1410.6337.63

Pattern of EQ-5D scores and pain visual analogue scale scores within the SAPPHIRE data set (for complete measurements at 12 months)

EQ-5D groupingnEQ-5D (mean)Pain VAS (mean)
< 012−0.0557.83
0–0.249130.155.77
0.25–0.49910.265
0.5–0.699500.6339.16
0.7–0.799300.7534.63
0.8–0.899310.8317.00
0.9–1.0341.05.29
Full index341.05.29
Overall1710.6329.99

Pattern of EQ-5D scores and pain visual analogue scale scores within the SAPPHIRE data set (for complete responses) at different time points

FIGURE 38. One-month follow-up.

FIGURE 38One-month follow-up

FIGURE 39. Three-month follow-up.

FIGURE 39Three-month follow-up

FIGURE 40. Twelve-month follow-up.

FIGURE 40Twelve-month follow-up

Regression models using pain visual analogue scale main effects with and without squared terms (using individual-level data at 1, 3 and 12 months)

Model 1Model 2Model 3Model 4
Econometric estimation methodOLSOLSTOBITCLAD
Independent variablesPain VAS nightPain VAS night, squared termsPain VAS night, squared termsPain VAS night, squared terms
Coeff.SECoeff.SECoeff.SECoeff.SE
Intercept0.77376890.00205010.7359170.08607940.801350.801350.80.0242477
pVASnight−0.00316190.0825151−0.00339670.0024525−0.0054838−0.0054838−0.00429170.0016361
pVASnight*pVASnight0.00000280.00001040.00001990.00001990.00002080.00002
R20.10090.1009
ME0.030510.030550.00183−0.04091
MAE0.201440.201670.194490.18155vw
RMSE0.264970.264870.264030.26689
|Differencel|a
|Δ| ≤ 0.106935%7036%7337%9046%
|Δ| ≤ 0.053920%4020%3920%5026%
|Δ| ≤ 0.0153%53%84%21%
ActualPredictPredictPredictPredict
Mean0.652500.621990.621950.650670.69341
SD0.2806350.0961520.0963830.1184350.080537
Min.−0.2400.4180.4240.4520.579
Max.1.000.7340.7360.8010.800
Rangeb1.2400.3160.3120.3490.221

SE, robust standard error.

a

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

b

Range = maximum value - minimum value.

Total n = 491. Estimation data set = 295 (60%), validation data set = 196.

Regression models using the main effects with and without squared terms and interaction term (using models estimated from 3-month data and used to predict EQ-5D scores at 12 months)

Model 1Model 2Model 3Model 4
Econometric estimation methodOLSOLSTOBITCLAD
Independent variablesPain VAS nightPain VAS night, squared termsPain VAS night, squared termsPain VAS night, squared terms
Coeff.SECoeff.SECoeff.SECoeff.SE
Intercept0.55408960.05245050.53537850.06639950.54139210.07255110.68726090.0621687
pVASnight0.00138530.00106720.00332830.00432930.00421450.00475310.00095220.0036473
pVASnight*pVASnight−0.0000230.0000497−0.00003120.0000546−0.0000130.0000385
R20.0081−0.0014
ME0.039580.04020.0206−0.0479
MAE0.217180.22180.21580.1877
RMSE0.2769510.2800380.2803510.279724
|Differencel|a
|Δ| ≤ 0.104124%4023%4828%6739%
|Δ| ≤ 0.051811%2112%2213%3621%
|Δ| ≤ 0.0153%21%74%2012%
ActualPredictPredictPredictPredict
Mean0.664090.595630.594420.612470.69277
SD0.3048620.0410600.0487730.0578190.009118
Min.−0.2400.5540.5350.5410.659
Max.1.0000.6870.6560.6840.705
Rangeb1.2400.1330.1200.1420.046

SE, robust standard error.

a

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

b

Range = maximum value - minimum value.

Validation data set = 171.

Actual versus predicted EQ-5D: mapping pain visual analogue scale to EQ-5D

FIGURE 41. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS1 model.

FIGURE 41Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS1 model

FIGURE 42. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS2 model.

FIGURE 42Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, OLS2 model

FIGURE 43. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, TOBIT model.

FIGURE 43Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, TOBIT model

FIGURE 44. Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, CLAD model.

FIGURE 44Observed and predicted EQ-5D scores using 1, 3 and 12 months' data, CLAD model

With adjustment for repeated measurements at 1, 3 and 12 months' follow-up.

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

Range = maximum value - minimum value.

With adjustment for repeated measurements at 1, 3 and 12 months' follow-up.

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

Range = maximum value - minimum value.

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

Range = maximum value - minimum value.

Proportion of predicted values with errors < |0.01|, |0.05| and |0.10|.

Range = maximum value - minimum value.

© 2012, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Cover of Management of Frozen Shoulder: A Systematic Review and Cost-Effectiveness Analysis
Management of Frozen Shoulder: A Systematic Review and Cost-Effectiveness Analysis.
Health Technology Assessment, No. 16.11.
Maund E, Craig D, Suekarran S, et al.

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...