1 / 18

Matched designs

Matched designs. Need Matched analysis. Incorrect unmatched analysis. . cc cc exp,exact Proportion | Exposed Unexposed | Total Exposed -----------------+------------------------+----------------------

jessicacantu
Download Presentation

Matched designs

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Matched designs Need Matched analysis

  2. Incorrect unmatched analysis • . cc cc exp,exact • Proportion • | Exposed Unexposed | Total Exposed • -----------------+------------------------+---------------------- • Cases | 106 515 | 621 0.1707 • Controls | 95 526 | 621 0.1530 • -----------------+------------------------+---------------------- • Total | 201 1041 | 1242 0.1618 • | | • | Point estimate | [95% Conf. Interval] • |------------------------+---------------------- • Odds ratio | 1.139622 | .8327338 1.560771 (exact) • Attr. frac. ex. | .122516 | -.200864 .359291 (exact) • Attr. frac. pop | .0209125 | • +----------------------------------------------- • 1-sided Fisher's exact P = 0.2205 • 2-sided Fisher's exact P = 0.4411 • This analysis ignores that a matching control was found for each case. Notice that the ‘sample size’ looks to be 1242 and yet nevertheless there is no evidence of a disease-exposure relationship.

  3. Correct classical analysis • . reshape wide cc exp,i(pair) j(ct) • (note: j = 1 2) • Data long -> wide • ----------------------------------------------------------------------------- • Number of obs. 1242 -> 621 • Number of variables 4 -> 5 • j variable (2 values) ct -> (dropped) • xij variables: • cc -> cc1 cc2 • exp -> exp1 exp2 • ----------------------------------------------------------------------------- • . mcc exp2 exp1 • | Controls | • Cases | Exposed Unexposed | Total • -----------------+------------------------+---------- • Exposed | 90 16 | 106 • Unexposed | 5 510 | 515 • -----------------+------------------------+---------- • Total | 95 526 | 621 • McNemar's chi2(1) = 5.76 Prob > chi2 = 0.0164 • Exact McNemar significance probability = 0.0266 • odds ratio 3.2 1.120172 11.16902 (exact). • The ‘sample size’ is 21! But the p-value is less than 5% and the estimated odds ratio is very different from the incorrect analysis

  4. Exact p-value is just the binomial • . bitesti 21 16 0.5 • N Observed k Expected k Assumed p Observed p • ------------------------------------------------------------ • 21 16 10.5 0.50000 0.76190 • Pr(k >= 16) = 0.013302 (one-sided test) • Pr(k <= 16) = 0.996401 (one-sided test) • Pr(k <= 5 or k >= 16) = 0.026604 (two-sided test)

  5. Conditional logistic regression version of the correct classical analysis • . clogit exp cc,group(pair) • note: multiple positive outcomes within groups encountered. • note: 600 groups (1200 obs) dropped due to all positive or • all negative outcomes. • Conditional (fixed-effects) logistic regression Number of obs = 42 • LR chi2(1) = 6.06 • ------------------------------------------------------------------------------ • exp | Coef. Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • cc | 1.163151 .5123475 2.27 0.023 .1589681 2.167334 • ------------------------------------------------------------------------------ • . clogit exp cc,group(pair) or • note: multiple positive outcomes within groups encountered. • note: 600 groups (1200 obs) dropped due to all positive or • all negative outcomes. • Conditional (fixed-effects) logistic regression Number of obs = 42 • LR chi2(1) = 6.06 • ------------------------------------------------------------------------------ • exp | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • cc | 3.2 1.639512 2.27 0.023 1.172301 8.734961 • P-values / CIs are based on the normal approximation to the binomial. • 600 concordant pairs are correctly ‘dropped’

  6. 4 matching controls per case • LA Study of endometrial cancer • use "C:\Mdsc643.02\la.dta", clear • (LA Study of Endometrial Cancer) • . desc • Contains data from C:\Mdsc643.02\la.dta • obs: 315 LA Study of Endometrial Cancer • vars: 11 24 Nov 1997 15:43 • size: 15,120 (98.6% of memory free) (_dta has notes) • ------------------------------------------------------------------------------- • storage display value • variable name type format label variable label • ------------------------------------------------------------------------------- • row float %9.0g • age float %9.0g Age (yr) • gbd float %9.0g yn Gall Bladder Disease • hyp float %9.0g yn Hyertension • obe float %9.0g yn Obesity • est float %9.0g yn Estrogen (Any) Use • conj float %9.0g cl Conjugated Dose • dur float %9.0g Estrogen Duration (mo) • ned float %9.0g yn Non Estrogen Drug • cc float %9.0g ccl Case/Control • quint float %9.0g 4 Controls: 1 Case • ------------------------------------------------------------------------------- • Sorted by: quint

  7. Incorrect analysis • . cc cc est,exact • Proportion • | Exposed Unexposed | Total Exposed • -----------------+------------------------+---------------------- • Cases | 56 7 | 63 0.8889 • Controls | 127 125 | 252 0.5040 • -----------------+------------------------+---------------------- • Total | 183 132 | 315 0.5810 • | | • | Point estimate | [95% Conf. Interval] • |------------------------+---------------------- • Odds ratio | 7.874016 | 3.38601 21.15736 (exact) • Attr. frac. ex. | .873 | .7046671 .9527351 (exact) • Attr. frac. pop | .776 | • +----------------------------------------------- • 1-sided Fisher's exact P = 0.0000 • 2-sided Fisher's exact P = 0.0000

  8. Classical analysis • . drop row • . sort quint cc • . by quint: gen otf=_n • . reshape wide cc age gbd hyp obe est conj dur ned , i(quint) j(otf) • (note: j = 1 2 3 4 5) • Data long -> wide • ----------------------------------------------------------------------------- • Number of obs. 315 -> 63 • Number of variables 11 -> 46 • j variable (5 values) otf -> (dropped) • xij variables: • cc -> cc1 cc2 ... cc5 • age -> age1 age2 ... age5 • gbd -> gbd1 gbd2 ... gbd5 • hyp -> hyp1 hyp2 ... hyp5 • obe -> obe1 obe2 ... obe5 • est -> est1 est2 ... est5 • conj -> conj1 conj2 ... conj5 • dur -> dur1 dur2 ... dur5 • ned -> ned1 ned2 ... ned5 • -----------------------------------------------------------------------------

  9. A new table • . gen sumcon=est1+est2+est3+est4 • . gen sumcas=est5 • . table sumcas sumcon • ---------------------------------------- • | sumcon • sumcas | 0 1 2 3 4 • ----------+----------------------------- • 0 | 4 1 11 • 1 | 317 1615 5 • ---------------------------------------- • There are 5 concordant pairs. • Exact p-values based on Binomial p= 1/5, 2/5, 3/5 and 4/5

  10. Components to the p-value • . bitesti 7 3 .2 • N Observed k Expected k Assumed p Observed p • ------------------------------------------------------------ • 7 3 1.4 0.20000 0.42857 • Pr(k >= 3) = 0.148032 (one-sided test) • Pr(k <= 3) = 0.966656 (one-sided test) • Pr(k >= 3) = 0.148032 (two-sided test) • note: lower tail of two-sided p-value is empty • . bitesti 18 17 .4 • N Observed k Expected k Assumed p Observed p • ------------------------------------------------------------ • 18 17 7.2 0.40000 0.94444 • Pr(k >= 17) = 0.000002 (one-sided test) • Pr(k <= 17) = 1.000000 (one-sided test) • Pr(k >= 17) = 0.000002 (two-sided test) • note: lower tail of two-sided p-value is empty • return list • scalars: • r(p) = 1.92414534861e-06

  11. Next 2 p-values • . bitesti 17 16 .6 • N Observed k Expected k Assumed p Observed p • ------------------------------------------------------------ • 17 16 10.2 0.60000 0.94118 • Pr(k >= 16) = 0.002088 (one-sided test) • Pr(k <= 16) = 0.999831 (one-sided test) • Pr(k <= 3 or k >= 16) = 0.002539 (two-sided test) • . bitesti 16 15 .8 • N Observed k Expected k Assumed p Observed p • ------------------------------------------------------------ • 16 15 12.8 0.80000 0.93750 • Pr(k >= 15) = 0.140737 (one-sided test) • Pr(k <= 15) = 0.971853 (one-sided test) • Pr(k <= 10 or k >= 15) = 0.222425 (two-sided test)

  12. Correct p-value • TITLE • STB-49 sbe28. Meta-analysis of p values. • DESCRIPTION/AUTHOR(S) • STB insert by Aurelio Tobias, Statistical Consultant, Madrid, Spain. • Support: bledatobias@ctv.es • After installation, see help metap. • INSTALLATION FILES (click here to install) • sbe28/metap.ado • sbe28/metap.hlp • ANCILLARY FILES (click here to get) • sbe28/fleiss.dta

  13. Using the STB ado file • . input pvar • pvar • 1. 0.148032 • 2. 1.92414534861e-06 • 3. 0.002539 • 4. 0.222425 • 5. end • . metap pvar • Meta-analysis of p_values • ------------------------------------------------------------ • Method | chi2 p_value studies • --------------------+--------------------------------------- • Fisher | 45.101012 3.521e-07 4 • ------------------------------------------------------------

  14. Conditional logistic version • . clogit est cc,group(quint) • note: multiple positive outcomes within groups encountered. • note: 5 groups (25 obs) dropped due to all positive or • all negative outcomes. • Conditional (fixed-effects) logistic regression Number of obs = 290 • LR chi2(1) = 35.35 • ------------------------------------------------------------------------------ • est | Coef. Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • cc | 2.07376 .4208244 4.93 0.000 1.248959 2.898561 • ------------------------------------------------------------------------------ • . clogit est cc,group(quint) or • note: multiple positive outcomes within groups encountered. • note: 5 groups (25 obs) dropped due to all positive or • all negative outcomes. • Conditional (fixed-effects) logistic regression Number of obs = 290 • LR chi2(1) = 35.35 • Prob > chi2 = 0.0000 • Log likelihood = -99.934552 Pseudo R2 = 0.1503 • ------------------------------------------------------------------------------ • est | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • cc | 7.954675 3.347522 4.93 0.000 3.486712 18.148 • ------------------------------------------------------------------------------

  15. Reversing Case/Control and Exposure • . clogit cc est,group(quint) • Conditional (fixed-effects) logistic regression Number of obs = 315 • LR chi2(1) = 35.35 • Prob > chi2 = 0.0000 • Log likelihood = -83.72159 Pseudo R2 = 0.1743 • ------------------------------------------------------------------------------b • cc | Coef. Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • est | 2.073761 .4208245 4.93 0.000 1.24896 2.898562 • ------------------------------------------------------------------------------ • . clogit cc est,group(quint) or • Conditional (fixed-effects) logistic regression Number of obs = 315 • LR chi2(1) = 35.35 • Prob > chi2 = 0.0000 • Log likelihood = -83.72159 Pseudo R2 = 0.1743 • ------------------------------------------------------------------------------ • cc | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • est | 7.954681 3.347525 4.93 0.000 3.486714 18.14802 • ------------------------------------------------------------------------------ • .

  16. Assessment of potential confounder • . clogit est hyp cc,group(quint) or • note: multiple positive outcomes within groups encountered. • note: 5 groups (25 obs) dropped due to all positive or • all negative outcomes. • Iteration 0: log likelihood = -93.816541 • Iteration 1: log likelihood = -93.775297 • Iteration 2: log likelihood = -93.775233 • Iteration 3: log likelihood = -93.775233 • Conditional (fixed-effects) logistic regression Number of obs = 290 • LR chi2(2) = 47.66 • Prob > chi2 = 0.0000 • Log likelihood = -93.775233 Pseudo R2 = 0.2026 • ------------------------------------------------------------------------------ • est | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • hyp | 3.175014 1.099597 3.34 0.001 1.610462 6.259518 • cc | 7.423919 3.149142 4.73 0.000 3.232684 17.04917 • ------------------------------------------------------------------------------

  17. Assessment of age as a potential modifier (even though age was a part of the matching criteria) • gen ac=age*cc • . clogit est hyp cc ac,group(quint) or • note: multiple positive outcomes within groups encountered. • note: 5 groups (25 obs) dropped due to all positive or • all negative outcomes. • Iteration 0: log likelihood = -94.0779 • Iteration 1: log likelihood = -93.779779 • Iteration 2: log likelihood = -93.774339 • Iteration 3: log likelihood = -93.774338 • Conditional (fixed-effects) logistic regression Number of obs = 290 • LR chi2(3) = 47.67 • Prob > chi2 = 0.0000 • Log likelihood = -93.774338 Pseudo R2 = 0.2027 • ------------------------------------------------------------------------------ • est | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval] • -------------+---------------------------------------------------------------- • hyp | 3.173201 1.100016 3.33 0.001 1.608502 6.259989 • cc | 6.085438 28.69545 0.38 0.702 .0005895 62816.24 • ac | 1.002775 .0656825 0.04 0.966 .8819609 1.140139 • ------------------------------------------------------------------------------

  18. Notice… • …that age*cc is included in the model even though age is not included. • This is a special case where we CAN interpret a model with an interaction term even though one of the constituents of this interaction is not included in the model

More Related