dat.gibson2002.RdResults from 15 trials examining the effectiveness of self-management education and regular medical review for adults with asthma.
dat.gibson2002The data frame contains the following columns:
| author | character | first author of study |
| year | numeric | publication year |
| n1i | numeric | number of participants in the intervention group |
| m1i | numeric | mean number of days off work/school in the intervention group |
| sd1i | numeric | standard deviation of the number of days off work/school in the intervention group |
| n2i | numeric | number of participants in the control/comparison group |
| m2i | numeric | mean number of days off work/school in the control/comparison group |
| sd2i | numeric | standard deviation of the number of days off work/school in the control/comparison group |
| ai | numeric | number of participants who had one or more days off work/school in the intervention group |
| bi | numeric | number of participants who no days off work/school in the intervention group |
| ci | numeric | number of participants who had one or more days off work/school in the control/comparison group |
| di | numeric | number of participants who no days off work/school in the control/comparison group |
| type | numeric | numeric code for the intervention type (see ‘Details’) |
Asthma management guidelines typically recommend for patients to receive education and regular medical review. While self-management programs have been shown to increase patient knowledge, it is less clear to what extent they actually impact health outcomes. The systematic review by Gibson et al. (2002) examined the effectiveness of self-management education and regular medical review for adults with asthma. In each study, participants receiving a certain management intervention were compared against those in a control/comparison group with respect to a variety of health outcomes. One of the outcomes examined in a number of studies was the number of days off work/school.
The majority of studies reporting this outcome provided means and standard deviations allowing a meta-analysis of standardized mean differences. Seven studies also reported the number of participants who had one or more days off work/school in each group. These studies could be meta-analyzed using, for example, (log) risk ratios. Finally, one could also consider a combined analysis based on standardized mean differences computed from the means and standard deviations where available and using probit transformed risk differences (which also provide estimates of the standardized mean difference) for the remaining studies.
Some degree of patient education was provided in all studies. In addition, the type variable indicates what additional intervention components were included in each study:
optimal self-management (writing action plan, self-monitoring, regular medical review),
self-monitoring and regular medical review,
self-monitoring only,
regular medical review only,
written action plan only.
Gibson, P. G., Powell, H., Wilson, A., Abramson, M. J., Haywood, P., Bauman, A., Hensley, M. J., Walters, E. H., & Roberts, J. J. L. (2002). Self-management education and regular practitioner review for adults with asthma. Cochrane Database of Systematic Reviews, 3, CD001117. https://doi.org/10.1002/14651858.CD001117
medicine, primary care, risk ratios, standardized mean differences
### copy data into 'dat' and examine data
dat <- dat.gibson2002
dat
#> author year n1i m1i sd1i n2i m2i sd2i ai bi ci di type
#> 1 Cote 1997 50 2.20 12.73 54 5.20 12.50 NA NA NA NA 1
#> 2 Ghosh 1998 140 17.60 24.20 136 34.10 38.80 NA NA NA NA 1
#> 3 Hayward 1996 23 0.38 0.56 19 0.23 0.29 NA NA NA NA 1
#> 4 Heard 1999 97 2.09 5.93 94 2.66 4.95 34 63 36 58 1
#> 5 Ignacio-Garcia 1995 35 4.92 6.05 35 20.00 26.34 24 11 29 6 1
#> 6 Knoell 1998 45 0.85 4.75 55 2.31 9.16 NA NA NA NA 1
#> 7 Lahdensuo 1996 56 2.80 9.00 59 4.80 7.20 13 43 25 34 1
#> 8 Sommaruga 1995 20 24.10 11.80 20 31.80 17.90 NA NA NA NA 1
#> 9 Zeiger 1991 128 1.40 3.30 143 2.30 7.60 NA NA NA NA 1
#> 10 Garret 1994 119 6.23 12.20 100 5.71 8.57 58 42 57 33 2
#> 11 Neri 1996 32 2.10 8.00 33 5.10 14.00 7 25 13 20 3
#> 12 Hilton 1986 86 0.73 1.48 100 0.47 1.20 NA NA NA NA 4
#> 13 Gallefoss 1999 25 8.00 32.00 24 26.00 70.00 NA NA NA NA 5
#> 14 Yoon 1993 28 NA NA 28 NA NA 5 23 4 24 1
#> 15 Brewin 1995 12 NA NA 33 NA NA 0 12 16 17 3
### load metafor package
library(metafor)
### compute standardized mean differences and corresponding sampling variances
dat <- escalc(measure="SMD", m1i=m1i, sd1i=sd1i, n1i=n1i, m2i=m2i, sd2i=sd2i, n2i=n2i, data=dat)
dat
#>
#> author year n1i m1i sd1i n2i m2i sd2i ai bi ci di type yi vi
#> 1 Cote 1997 50 2.20 12.73 54 5.20 12.50 NA NA NA NA 1 -0.2361 0.0388
#> 2 Ghosh 1998 140 17.60 24.20 136 34.10 38.80 NA NA NA NA 1 -0.5105 0.0150
#> 3 Hayward 1996 23 0.38 0.56 19 0.23 0.29 NA NA NA NA 1 0.3209 0.0973
#> 4 Heard 1999 97 2.09 5.93 94 2.66 4.95 34 63 36 58 1 -0.1038 0.0210
#> 5 Ignacio-Garcia 1995 35 4.92 6.05 35 20.00 26.34 24 11 29 6 1 -0.7804 0.0615
#> 6 Knoell 1998 45 0.85 4.75 55 2.31 9.16 NA NA NA NA 1 -0.1930 0.0406
#> 7 Lahdensuo 1996 56 2.80 9.00 59 4.80 7.20 13 43 25 34 1 -0.2445 0.0351
#> 8 Sommaruga 1995 20 24.10 11.80 20 31.80 17.90 NA NA NA NA 1 -0.4978 0.1031
#> 9 Zeiger 1991 128 1.40 3.30 143 2.30 7.60 NA NA NA NA 1 -0.1504 0.0148
#> 10 Garret 1994 119 6.23 12.20 100 5.71 8.57 58 42 57 33 2 0.0484 0.0184
#> 11 Neri 1996 32 2.10 8.00 33 5.10 14.00 7 25 13 20 3 -0.2589 0.0621
#> 12 Hilton 1986 86 0.73 1.48 100 0.47 1.20 NA NA NA NA 4 0.1937 0.0217
#> 13 Gallefoss 1999 25 8.00 32.00 24 26.00 70.00 NA NA NA NA 5 -0.3277 0.0828
#> 14 Yoon 1993 28 NA NA 28 NA NA 5 23 4 24 1 NA NA
#> 15 Brewin 1995 12 NA NA 33 NA NA 0 12 16 17 3 NA NA
#>
### fit an equal-effects model to the standardized mean differences (as in Gibson et al., 2002)
res <- rma(yi, vi, data=dat, method="EE")
#> Warning: 2 studies with NAs omitted from model fitting.
print(res, digits=2)
#>
#> Equal-Effects Model (k = 13)
#>
#> I^2 (total heterogeneity / total variability): 55.36%
#> H^2 (total variability / sampling variability): 2.24
#>
#> Test for Heterogeneity:
#> Q(df = 12) = 26.88, p-val < .01
#>
#> Model Results:
#>
#> estimate se zval pval ci.lb ci.ub
#> -0.18 0.05 -3.77 <.01 -0.28 -0.09 ***
#>
#> ---
#> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#>
### compute log risk ratios and corresponding sampling variances
dat <- escalc(measure="RR", ai=ai, bi=bi, ci=ci, di=di, data=dat)
dat
#>
#> author year n1i m1i sd1i n2i m2i sd2i ai bi ci di type yi vi
#> 1 Cote 1997 50 2.20 12.73 54 5.20 12.50 NA NA NA NA 1 NA NA
#> 2 Ghosh 1998 140 17.60 24.20 136 34.10 38.80 NA NA NA NA 1 NA NA
#> 3 Hayward 1996 23 0.38 0.56 19 0.23 0.29 NA NA NA NA 1 NA NA
#> 4 Heard 1999 97 2.09 5.93 94 2.66 4.95 34 63 36 58 1 -0.0886 0.0362
#> 5 Ignacio-Garcia 1995 35 4.92 6.05 35 20.00 26.34 24 11 29 6 1 -0.1892 0.0190
#> 6 Knoell 1998 45 0.85 4.75 55 2.31 9.16 NA NA NA NA 1 NA NA
#> 7 Lahdensuo 1996 56 2.80 9.00 59 4.80 7.20 13 43 25 34 1 -0.6017 0.0821
#> 8 Sommaruga 1995 20 24.10 11.80 20 31.80 17.90 NA NA NA NA 1 NA NA
#> 9 Zeiger 1991 128 1.40 3.30 143 2.30 7.60 NA NA NA NA 1 NA NA
#> 10 Garret 1994 119 6.23 12.20 100 5.71 8.57 58 42 57 33 2 -0.0880 0.0137
#> 11 Neri 1996 32 2.10 8.00 33 5.10 14.00 7 25 13 20 3 -0.5883 0.1582
#> 12 Hilton 1986 86 0.73 1.48 100 0.47 1.20 NA NA NA NA 4 NA NA
#> 13 Gallefoss 1999 25 8.00 32.00 24 26.00 70.00 NA NA NA NA 5 NA NA
#> 14 Yoon 1993 28 NA NA 28 NA NA 5 23 4 24 1 0.2231 0.3786
#> 15 Brewin 1995 12 NA NA 33 NA NA 0 12 16 17 3 -2.5351 1.9543
#>
### fit an equal-effects model to the log risk ratios
res <- rma(yi, vi, data=dat, method="EE")
#> Warning: 8 studies with NAs omitted from model fitting.
print(res, digits=2)
#>
#> Equal-Effects Model (k = 7)
#>
#> I^2 (total heterogeneity / total variability): 18.22%
#> H^2 (total variability / sampling variability): 1.22
#>
#> Test for Heterogeneity:
#> Q(df = 6) = 7.34, p-val = 0.29
#>
#> Model Results:
#>
#> estimate se zval pval ci.lb ci.ub
#> -0.17 0.08 -2.31 0.02 -0.32 -0.03 *
#>
#> ---
#> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#>
predict(res, transf=exp, digits=2)
#>
#> pred ci.lb ci.ub
#> 0.84 0.72 0.97
#>
### note: Gibson et al. (2002) used the Mantel-Haenszel method for their analysis
rma.mh(measure="RR", ai=ai, bi=bi, ci=ci, di=di, data=dat, digits=2)
#> Warning: 8 studies with NAs omitted from model fitting.
#> Warning: Some yi/vi values are NA.
#>
#> Equal-Effects Model (k = 7)
#>
#> I^2 (total heterogeneity / total variability): 25.74%
#> H^2 (total variability / sampling variability): 1.35
#>
#> Test for Heterogeneity:
#> Q(df = 6) = 8.08, p-val = 0.23
#>
#> Model Results (log scale):
#>
#> estimate se zval pval ci.lb ci.ub
#> -0.24 0.08 -2.91 <.01 -0.40 -0.08
#>
#> Model Results (RR scale):
#>
#> estimate ci.lb ci.ub
#> 0.79 0.67 0.92
#>
### compute standardized mean differences where possible and otherwise probit transformed
### risk differences (which also provide estimates of the standardized mean differences)
dat <- escalc(measure="SMD", m1i=m1i, sd1i=sd1i, n1i=n1i,
m2i=m2i, sd2i=sd2i, n2i=n2i, data=dat, add.measure=TRUE)
dat <- escalc(measure="PBIT", ai=ai, bi=bi, ci=ci, di=di, data=dat, replace=FALSE, add.measure=TRUE)
dat
#>
#> author year n1i m1i sd1i n2i m2i sd2i ai bi ci di type yi vi measure
#> 1 Cote 1997 50 2.20 12.73 54 5.20 12.50 NA NA NA NA 1 -0.2361 0.0388 SMD
#> 2 Ghosh 1998 140 17.60 24.20 136 34.10 38.80 NA NA NA NA 1 -0.5105 0.0150 SMD
#> 3 Hayward 1996 23 0.38 0.56 19 0.23 0.29 NA NA NA NA 1 0.3209 0.0973 SMD
#> 4 Heard 1999 97 2.09 5.93 94 2.66 4.95 34 63 36 58 1 -0.1038 0.0210 SMD
#> 5 Ignacio-Garcia 1995 35 4.92 6.05 35 20.00 26.34 24 11 29 6 1 -0.7804 0.0615 SMD
#> 6 Knoell 1998 45 0.85 4.75 55 2.31 9.16 NA NA NA NA 1 -0.1930 0.0406 SMD
#> 7 Lahdensuo 1996 56 2.80 9.00 59 4.80 7.20 13 43 25 34 1 -0.2445 0.0351 SMD
#> 8 Sommaruga 1995 20 24.10 11.80 20 31.80 17.90 NA NA NA NA 1 -0.4978 0.1031 SMD
#> 9 Zeiger 1991 128 1.40 3.30 143 2.30 7.60 NA NA NA NA 1 -0.1504 0.0148 SMD
#> 10 Garret 1994 119 6.23 12.20 100 5.71 8.57 58 42 57 33 2 0.0484 0.0184 SMD
#> 11 Neri 1996 32 2.10 8.00 33 5.10 14.00 7 25 13 20 3 -0.2589 0.0621 SMD
#> 12 Hilton 1986 86 0.73 1.48 100 0.47 1.20 NA NA NA NA 4 0.1937 0.0217 SMD
#> 13 Gallefoss 1999 25 8.00 32.00 24 26.00 70.00 NA NA NA NA 5 -0.3277 0.0828 SMD
#> 14 Yoon 1993 28 NA NA 28 NA NA 5 23 4 24 1 0.1467 0.1627 PBIT
#> 15 Brewin 1995 12 NA NA 33 NA NA 0 12 16 17 3 -1.7320 0.4546 PBIT
#>
### fit a random-effects model to these estimates
res <- rma(yi, vi, data=dat)
print(res, digits=2)
#>
#> Random-Effects Model (k = 15; tau^2 estimator: REML)
#>
#> tau^2 (estimated amount of total heterogeneity): 0.04 (SE = 0.03)
#> tau (square root of estimated tau^2 value): 0.20
#> I^2 (total heterogeneity / total variability): 52.75%
#> H^2 (total variability / sampling variability): 2.12
#>
#> Test for Heterogeneity:
#> Q(df = 14) = 32.82, p-val < .01
#>
#> Model Results:
#>
#> estimate se zval pval ci.lb ci.ub
#> -0.20 0.08 -2.67 <.01 -0.35 -0.05 **
#>
#> ---
#> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#>
### meta-regression model examining if there are systematic differences based on the
### type of measure used (there are only 2 studies where measure="PBIT", so this isn't
### very conclusive here, but shown for illustration purposes)
res <- rma(yi, vi, mods = ~ measure, data=dat)
print(res, digits=2)
#>
#> Mixed-Effects Model (k = 15; tau^2 estimator: REML)
#>
#> tau^2 (estimated amount of residual heterogeneity): 0.04 (SE = 0.03)
#> tau (square root of estimated tau^2 value): 0.21
#> I^2 (residual heterogeneity / unaccounted variability): 56.02%
#> H^2 (unaccounted variability / sampling variability): 2.27
#> R^2 (amount of heterogeneity accounted for): 0.00%
#>
#> Test for Residual Heterogeneity:
#> QE(df = 13) = 32.60, p-val < .01
#>
#> Test of Moderators (coefficient 2):
#> QM(df = 1) = 0.28, p-val = 0.59
#>
#> Model Results:
#>
#> estimate se zval pval ci.lb ci.ub
#> intrcpt -0.40 0.38 -1.06 0.29 -1.15 0.34
#> measureSMD 0.21 0.39 0.53 0.59 -0.56 0.97
#>
#> ---
#> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#>
predict(res, newmods=1, digits=2)
#>
#> pred se ci.lb ci.ub pi.lb pi.ub
#> -0.20 0.08 -0.35 -0.04 -0.63 0.24
#>