The DHS Program User Forum: Dataset access » Restructuring Healthcare data

Home » Data » Dataset access » Restructuring Healthcare data

Show: Today's Messages :: Show Polls :: Message Navigator

Restructuring Healthcare data [message #7104]

Mon, 24 August 2015 03:13

JREY
Messages: 13
Registered: August 2015

Member

Hi, everyone. I am examining health care utilization in the Philippines using the data using the NDHS household questionnaire. In the questionnaire the household respondent was asked if there was any member of the household who visited a health facility for consultation/advice or treatment anywhere. For each member who visited a health facility, the reason for consultation and the type of treatment received was asked. I would like to create a dataset where each household member has an information on these health care indicators, but I find it hard to structure the data in order to come up with my desired dataset. Could anyone please help me on this? Thank you.

Report message to a moderator

Re: Restructuring Healthcare data [message #8111 is a reply to message #7104]

Tue, 25 August 2015 07:48

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Are you a Stata user?

Report message to a moderator

Re: Restructuring Healthcare data [message #8112 is a reply to message #8111]

Tue, 25 August 2015 08:05

JREY
Messages: 13
Registered: August 2015

Member

I use STATA and SPSS.

Report message to a moderator

Re: Restructuring Healthcare data [message #8118 is a reply to message #8112]

Tue, 25 August 2015 12:26

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is a response from Senior DHS Stata Specialist, Tom Pullum:

I can understand why you are having difficulty. I see that this information was not coded onto the PR file in an efficient way. Every record in the household has all the information about the visits of all the members. It would have been better to code the information about, say, the person on line 3, just onto line 3. I think the best way to deal with this is to separate the information you need out, organize it by line number, and then merge back with the PR file.

The attached program will do this. It is a text file; you will probably want to put it into the Stata do file editor. You will need to change the paths. You will want to add more variables; I have shown where to do this.

The program puts the health information about each person onto that person's record. The file will only include the specific people who have had a visit. Let me know if you have questions.

Attachment: PH_health_do_25Aug2015.txt
(Size: 2.42KB, Downloaded 911 times)

Report message to a moderator

Re: Restructuring Healthcare data [message #8121 is a reply to message #8118]

Wed, 26 August 2015 04:22

JREY
Messages: 13
Registered: August 2015

Member

Thank you very much. This is very useful!

Report message to a moderator

Re: Restructuring Healthcare data [message #8123 is a reply to message #8121]

Wed, 26 August 2015 06:55

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Also from Tom Pillum:

I realized that there is a simpler and more direct way to do the same thing that was described in my previous response. This Stata program will also put the variables sh209, sh210, etc., onto the record for the correct person in the household, but without requiring a merge. The result should be the same.

Attachment: PH_health_do_25Aug2015_v2.txt
(Size: 1.64KB, Downloaded 910 times)

[Updated on: Wed, 11 November 2015 11:47]

Report message to a moderator

Re: Restructuring Healthcare data [message #8131 is a reply to message #8123]

Thu, 27 August 2015 05:00

JREY
Messages: 13
Registered: August 2015

Member

Thanks again for this command file.

I am also interested in creating a similar file for each member in the household who were sick/injured in the last 30 days.I am using the command that you sent but replaced the variables with the relevant ones. However, I got stuck in the looping command after I generated the seqno variable. I changed the sequence number to 12 and the person's line number to sh203, but it is still not working. Could you please help me on this?

Thank you.

Report message to a moderator

Re: Restructuring Healthcare data [message #8149 is a reply to message #8131]

Fri, 28 August 2015 13:38

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Another response from Tom Pullum:

Please try the attached program. The main problem is that some of the indexed variables, such as sh203, have indices with a leading 0--for example, 01, 02, 03, rather than 1, 2, 3. Stata has a very powerful rename command, so you need to include " rename *_0* *_* " to get rid of those 0's.

Hope this will work.

Attachment: PH_sick_injured_do_27Aug2015.txt
(Size: 2.12KB, Downloaded 971 times)

Report message to a moderator

Re: Restructuring Healthcare data [message #8160 is a reply to message #8149]

Mon, 31 August 2015 04:44

JREY
Messages: 13
Registered: August 2015

Member

Thank you very much for your help!

Does it make sense to merge the "sick/injured" file and the "healthcare utilization" file into one dataset using the hv001 hv002 and hvidx as the ID?

Again thank you very much!

Report message to a moderator

Re: Restructuring Healthcare data [message #8230 is a reply to message #8160]

Fri, 18 September 2015 14:15

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Another response from Tom Pullum:

Yes, if you do the kind of reshaping I suggested earlier for both, then you can merge them back together. It would also be possible to reshape both types of variables within a single program. Either way would work.

Report message to a moderator