Network meta-analysis

[Paola Caro]

Hello. I have the following data set.
I get the 'variables study_id treatment do not uniquely identify the observations' every time I run network setup n mean sd, studyvar(study_id) trtvar(treatment). Any idea what I am doing wrong when entering the data? an how to deal with treatments that have different arms?

No	study_id	treatment	Arm	n	mean	sd	time_point	outcome
85	Celano 1996	Conventional treatment		16	5.6	4.9	Posttreatment	PTSD
86	Celano 1996	CBT		14	4.8	3.2	Posttreatment	PTSD
89	Celano 1996	Conventional treatment		17	41.8	8.4	Posttreatment	PTSD
90	Celano 1996	CBT		15	39.9	9.6	Posttreatment	PTSD
177	Cohen 2005	CBT		41	8.78	4.88	Posttreatment	PTSD
178	Cohen 2005	Child Centered Therapy		41	9.92	5.28	Posttreatment	PTSD
309	Danielson 2012	Risk Reduction Family Therapy		15	17.9	12.8	Posttreatment	PTSD
310	Danielson 2012	Conventional treatment		15	22.9	13.5	Posttreatment	PTSD
415	Deblinger 1999	CBT	Mother only	20	5	3.37	Posttreatment	PTSD
416	Deblinger 1999	CBT	Child Only	21	3.38	3.02	Posttreatment	PTSD
417	Deblinger 1999	CBT	Mother and child	19	3.68	3.07	Posttreatment	PTSD
418	Deblinger 1999	Conventional treatment		14	6.5	3.39	Posttreatment	PTSD
485	Deblinger 2001	Supprtive therapy		23	12.35	11.67	Posttreatment	PTSD
486	Deblinger 2001	CBT		21	14.67	20.28	Posttreatment	PTSD
503	Deblinger 2001	Supprtive therapy		23	6.09	6.73	Posttreatment	PTSD
504	Deblinger 2001	CBT		21	6.57	7.92	Posttreatment	PTSD
621	Foa 2013	CBT		31	6.7	19.09	Posttreatment	PTSD
622	Foa 2013	Child Centered Therapy		30	16.1	19.28	Posttreatment	PTSD
627	Foa 2013	CBT		31	9.9	20.72	Posttreatment	PTSD
628	Foa 2013	Child Centered Therapy		30	16.1	20.62	Posttreatment	PTSD
645	Jaberghaderi 2004	EMDR		7	18.86	7.9	Posttreatment	PTSD
646	Jaberghaderi 2004	CBT		7	22.71	6.9	Posttreatment	PTSD
649	Jaberghaderi 2004	EMDR		7	10.14	5.4	Posttreatment	PTSD
650	Jaberghaderi 2004	CBT		7	11.29	6.6	Posttreatment	PTSD
655	Kaczkurkin 2016	CBT		31	7.69	8.34	Posttreatment	PTSD
656	Kaczkurkin 2016	Child Centered Therapy		30	15.63	10.77	Posttreatment	PTSD
666	King 2000	CBT	Child	9	7.58	4.42	Posttreatment	PTSD
667	King 2000	CBT	Family	9	6.5	5.14	Posttreatment	PTSD
668	King 2000	Waiting list		10	11.36	2.11	Posttreatment	PTSD
693	King 2000	CBT	Child	9	14.72	7.7	Posttreatment	PTSD
694	King 2000	CBT	Family	9	11.25	4.99	Posttreatment	PTSD
695	King 2000	Waiting list		10	17.3	8.31	Posttreatment	PTSD
728	McLean 2017	CBT		31	22.03	10.95	Posttreatment	PTSD
729	McLean 2017	Child Centered Therapy		30	15.42	11.1	Posttreatment	PTSD
736	Farkas 2008	MASTR-EMDR		15	2.83	3.34	Posttreatment	PTSD
737	Farkas 2008	Waiting list		15	10.13	7.39	Posttreatment	PTSD

Thank you a lot for any help.

[Marcos Almeida]

The display is not under - dataex - hence it is not so easy to read. That said, it seems that some study IDs + treatments are repeated. If they convey subgroups or substudies, you may add, say "b" to the study ID, for example.

[Paola Caro]

Hi Marcos.
Thank you for your time and help.
The studies and treatments id's are repeated because I have some studies measuring the same outcome with different instruments and also some studies having four arms. For example: three types of CBT vs Waiting list. Apologies. What do you mean by adding a b? would it still work after explaining the repeated values? How can I paste better the data?
Thank you again!!
Paola

[Marcos Almeida]

I meant, for example: StudyDrBrown_a, StudyDrBown_b, then you could explain that you separated according to the instruments.

[Paola Caro]

Hello Marcos,

I tried, but still getting the same error. This is how I edited the data. Do you have any additional recommendation? Thank you very much indeed,

[Marcos Almeida]

Sorry, but this is the furtherst I can go. I did some network meta-analysis last year, I had a similar issue, and I 'll try to figure out how I curbed it as soon as I get some free time.

Meanwhile, hopefully you'll get insightful advice..

[Paola Caro]

Hello Marcos,

Thank you very much for any help. Very much appreciated.

Paola