link.springer.com10.1007/s11606... · web viewhemorrhoid-specific questions (de novo) sf-12v2 (...
TRANSCRIPT
ONLINE APPENDIX: EFFECT OF PERINEAL SELF-ACUPRESSURE ON
CONSTIPATION: A RANDOMIZED CONTROLLED TRIAL
Ryan Abbott MD, JD, MTOM12*
Ian Ayres PhD, JD3
Ed Hui MD1
Ka-Kit Hui MD1
1 Division of General Internal Medicine and Health Services Research, David Geffen School of
Medicine at University of California, Los Angeles, Department of Medicine
2 Southwestern Law School
3 Yale Law School
* Corresponding Author: Ryan Abbott, Department of Medicine, University of California, Los
Angeles, 1033 Gayley Ave, Suite 111, Los Angeles, CA 90024. Telephone: (310) 794-0712. Fax:
(310) 794-3310. Email: [email protected].
Online Appendix 1. Extended Literature Review
This supplementary material provides additional evidence from the literature on the specific mechanisms by which perineal pressure might aid in defecation.
Online Appendix 2. Patient Educational Materials
A. Perineal Self-Acupressure (Female Version)B. Perineal Self-Acupressure (Male Version) C. Patient Information: Constipation in Adults (The Basics) (© UpToDate)D. Patient Information: Hemorrhoids (The Basics) (© UpToDate)
Online Appendix 3. Patient Surveys
A. PAC-QOL (© 2001 Johnson & Johnson.)B. Modified Bowel Function IndexC. Hemorrhoid-specific questions (de novo)D. SF-12v2 ( 1994, 2002 Medical Outcomes Trust and QualityMetric Incorporated.)E. Treatment group only questionnaire (only asked after trial period, de novo)
1
Online Appendix 4. Supplementary Statistical Analyses
A. LOCF (all individual questions).B. Complete case analysis.C. Multiple imputation analysis.D. Seemingly unrelated regression analysis (with LOCF dataset).E. Ordered probit analysis (with LOCF dataset).F. Treatment group only questionnaire analysis. G. Instrumental variables analysis (with LOCF dataset).H. Demographics interaction analysis (with LOCF dataset).I. Pre-treatment symptom interaction analysis (with LOCF dataset).
Online Appendix 5. Statistical Code
This supplementary material has been provided by the authors to give readers additional information about their work.
2
Online Appendix 1
Extended Literature Review
Several studies have examined the use perineal pressure or transvaginal pressure to aid in defecation [12, “Many females with obstructed defecation apply digital pressure on their perineum to facilitate defecation.”; 12, 92% of patients with obstructed defecation “stated that they applied perineal pressure on a regular basis to facilitate their defecation” (p. 370); 31, “patients in whom constipation is relieved with digital vaginal pressure to facilitate defecation” (Chapter 18, pp. 8-18)]. Digitally applying transvaginal (or perineal) pressure has been used by patients with rectocele or descending perineum syndrome [31, “Women may report the need to use their thumb or fingers to support the posterior vaginal wall to complete defecation” (Chapter 18); 9; 12].
Perineal pressure has also been shown to substantially increase “rectal tone” (measured by decreased rectal balloon volume) [11, 12]. For example, the following figure taken from [12] shows the perineorectal reflex (which across 17 subjects produced a tonic response that averaged 52% greater than when subjects did not apply perineal pressure):
Source: Gosselink M, Schouten W (2002) [2002]
The literature suggests multiple mechanisms by which perineal or transvaginal pressure might aid defecation:
1) Break up scybalous stools, [13; 14, “In women, applying transvaginal pressure with the other hand may also aid fragmentation and expulsion” (p. 661)]; 2) Protect the tissue from overstretching and relax the anal sphincters to allow stool passage, [2]; 3) Relax the puborectalis allowing the anorectal angle to straighten, allowing a less obstructed descent of the stool [15; 16; 12, “Preston and Lennard-Jones have suggested that the puborectalis muscle is pushed upward by this maneuver. This might straighten the anorectal angle, enhancing the alignment between the rectum and anal canal” (p. 370)];4) Stimulate extrinsic parasympathetic sacral nerves (and pons cerebri) to enhance rectal sensory perception and rectal tone, [12]; 5) Compensate for rectal wall abnormalities (such as rectocele or descending perineum syndrome) [8, 9].
3
Just as impacted stools are fragmented by internal digital manipulation, it is possible that similar digital pressure applied to the perineum might be sufficient to pulverize hardened stool. Physiologically, because the stool during attempted defecation descends into the rectum and can be lodged between the internal sphincter and the pelvic diaphragm, the stool may be sufficiently stable to prevent dorsal or anterior movement of the stool in response to the perineal or transvaginal pressure. Moreover, because the perineum descends by 1.0 to 3.5 cm. during defecation, the posterior portion of the stool becomes more susceptible to dorsal pressure. Fragmenting or pulverizing the posterior portion is particularly important with slow-transit constipation, because the posterior portion is likely to be the most scybalous portion of the stool [26-29]. If needed, digital perineal pressure to fragment the stool can be applied repeatedly as soften portions of the stool are expulsed from the body.
In addition to treating constipation, perineal massage may treat hemorrhoids, which are related to chronic constipation and are associated excessive straining [2]. Symptoms of hemorrhoids include itching, pain, bleeding, rectal prolapse and vein thromboses. One important way to prevent the development and progression of hemorrhoids is to effectively treat constipation [2, 17]. Stool softeners and laxatives have been shown to be effective in reducing risk of hemorrhoid recurrence [17, 30]. One of the standard treatments for impacted stools is to manually disimpact the stool [13, 14].
References
[26] Cowgill GR, Anderson WE, Sullivan AJ. The form of the stool as a criterion of laxation. JAMA. 1933 Jul 22;101(4):273–5.
[27] Burnett, FL. Fecal Units and Intestinal Rate: A Basis for the Study of Health and Intestinal Indigestion. Boston Med Surg J. 1921 Apr 14;184:371–6.
[28] Burnett, FL. The Intestinal Rate and the Form of the Feces. Am J Roentgenol. 1923 Aug;10:599–604.
[29] Sonnenberg A, Koch TR. Physician visits in the United States for constipation: 1958 to 1986. Dig Dis Sci. 1989 Apr;34(4):606–11.
[30] Petticrew M, Rodgers M, Booth A. Effectiveness of laxatives in adults. Qual Health Care. 2001 Dec;10(4):268–73.
[31] Feldman M, Friedman LS, Brandt LJ, editors. 9th ed. Vol. 1. Sleisenger and Fordtran’s Gastrointestinal and Liver Disease: Pathophysiology/Diagnosis/ Management. Philadelphia: Saunders; c2010. 3184 p.
4
eAppendix 2
Below are the 4 handouts participants received. They are, in order:
A. Perineal Self-Acupressure (Female Version) (only given to female treatment group participants)B. Perineal Self-Acupressure (Male Version) (only given to male treatment group participants)C. Patient Information: Constipation in Adults (The Basics) (© UpToDate) (given to all
participants)D. Patient Information: Hemorrhoids (The Basics) (© UpToDate) (given to all participants)
Handout A (Female Version)
5
6
Handout B (Male Version)
7
Handout C
8
Handout D
9
10
Online Appendix 3
Below are the 5 surveys participants completed. They are, in order:
A. PAC-QOL (© 2001 Johnson & Johnson)B. Modified Bowel Function IndexC. Hemorrhoid-specific questions (de novo)D. SF-12v2 ( 1994, 2002 Medical Outcomes Trust and QualityMetric Incorporated)E. Treatment group only questionnaire (only asked after trial period, de novo)
11
Survey A
PAC-QOL
PATIENT ASSESSMENT OF CONSTIPATION
The following questions are designed to measure the impact constipation has had on your daily life over the past 2 weeks. For each question, please check one box.
The following questions ask about your symptoms related to constipation. During the past 2 weeks, to what extent or intensity have you...
Not at all0
A little bit1
Moderately2
Quite a bit3
Extremely4
1. felt bloated to the point of bursting?
2. felt heavy because of your constipation?
The next few questions ask about how constipation affects your daily life. During the past 2 weeks, how much of the time have you...
None of the time
0
A little ofthe time
1
Some ofthe time
2
Most ofthe time
3
All ofthe time
4
3. felt any physical discomfort?
4. felt the need to have a bowel movement but not been able to?
5. been embarrassed to be with other people?
6. been eating less and less because of not being able to have bowel movements?
12
The next few questions ask about how constipation affects your daily life. During the past 2 weeks, to what extent or intensity have you...
Not at all0
A little bit1
Moderately2
Quite a bit3
Extremely4
7. had to be careful about what you eat?
8. had a decreased appetite?
9. been worried about not being able to choose what you eat (for example, at a friend’s house)?
10. been embarrassed about staying in the bathroom for so long when you were away from home?
11. been embarrassed about having to go to the bathroom so often when you were away from home?
12. been worried about having to change your daily routine (for example, traveling, being away from home)?
The next few questions ask about your feelings related to constipation. During the past 2 weeks, how much of the time have you...
None ofthe time
0
A little ofthe time
1
Some ofthe time
2
Most ofthe time
3
All of the time
4
13. felt irritable because of your condition?
14. been upset by your condition?
15. felt obsessed by your condition?
16. felt stressed by your condition?
17. felt less self-confident because of your condition?
18. felt in control of your situation?
13
The next questions ask about your feelings related to constipation. During the past 2 weeks, to what extent or intensity have you...
Not at all0
A little bit1
Moderately2
Quite a bit3
Extremely4
19. been worried about not knowing when you are going to be able to have a bowel movement?
20. been worried about not being able to have a bowel movement?
21. been more and more bothered by not being able to have a bowel movement?
The next questions ask about your life with constipation. During the past 2 weeks, how much of the time have you...
None of the time
0
A little of the time
1
Some of the time
2
Most of the time
3
All of the time
4
22. been worried that your condition will get worse?
23. felt that your body was not working properly?
24. had fewer bowel movements than you would like?
The next questions ask about your degree of satisfaction related to constipation. During the past 2 weeks, to what extent or intensity have you been...
Not at all0
A little bit1
Moderately2
Quite a bit3
Extremely4
25. satisfied with how often you have a bowel movement?
26. satisfied with the regularity of your bowel movements?
27. satisfied with the time it takes for food to pass through the intestines?
28. satisfied with your treatment?
14
Survey B
MODIFIED BOWEL FUNCTION INDEX
1. During the last 7 days, how would you rate your ease of defecation on a scale from 0 to 100, where 0 = easy or no difficulty and 100 = severe difficulty?”
0 = easy / no difficulty100 = severe difficulty
__________________(0 – 100)
2. During the last 7 days, how would you rate your feeling of incomplete bowel evacuation on a scale from 0 to 100, where 0 = no feeling of incomplete evacuation and 100 = a very strong feeling of incomplete evacuation?
0 = not at all100 = very strong
__________________(0 – 100)
3. During the last 7 days, how would you rate your constipation on a scale from 0 to 100, where 0 = not at all and 100 = very strong?
0 = not at all100 = very strong
__________________(0 – 100)
15
Survey C
Hemorrhoid-specific questions
In the last 4 weeks, have you experienced new episodes with hemorrhoids?Yes ⎕ No ⎕
During the past 4 weeks, to what extent have you experienced the following symptoms associated with hemorrhoids
Not at all0
A little bit1
Moderately2
Quite a bit
3
Extremely
4Bleeding ⎕ ⎕ ⎕ ⎕ ⎕
Itching ⎕ ⎕ ⎕ ⎕ ⎕
Pain ⎕ ⎕ ⎕ ⎕ ⎕
16
Survey D
SF-12v2
Your Health and Well-Being
This survey asks for your views about your health. This information will help keep track of how you feel and how well you are able to do your usual activities. Thank you for completing this survey!
For each of the following questions, please mark an in the one box that best describes your answer.
1. In general, would you say your health is:
Excellent Very good Good Fair Poor
1 2 3 4 5
2. The following questions are about activities you might do during a typical day. Does your health now limit you in these activities? If so, how much?
Yes,limited
a lot
Yes,limiteda little
No, notlimitedat all
17
a Moderate activities, such as moving a table, pushing a vacuum cleaner, bowling, or playing golf................................................. 1................. 2................ 3
b Climbing several flights of stairs................................................................. 1................. 2................ 33.During the past 4 weeks , how much of the time have you had any of the following problems with your work or
other regular daily activities as a result of your physical health?
4.During the past 4 weeks , how much of the time have you had any of the following problems with your work or other regular daily activities as a result of any emotional problems (such as feeling depressed or anxious)?
5.During the past 4 weeks , how much did pain interfere with your normal work (including both work outside the home and housework)?
Not at all A little bit Moderately Quite a bit Extremely
1 2 3 4 5
18
All ofthe time
Most ofthe time
Some ofthe time
A little ofthe time
None ofthe time
a Accomplished less than you
would like................................................. 1................. 2................. 3.................. 4................ 5
b Were limited in the kind ofwork or other activities............................. 1................. 2................. 3.................. 4................ 5
All ofthe time
Most ofthe time
Some ofthe time
A little ofthe time
None ofthe time
a Accomplished less than you
would like................................................. 1................. 2................. 3.................. 4................ 5
b Did work or other activitiesless carefully than usual............................ 1................. 2................. 3.................. 4................ 5
6. These questions are about how you feel and how things have been with you during the past 4 weeks. For each question, please give the one answer that comes closest to the way you have been feeling. How much of the time during the past 4 weeks…
7.During the past 4 weeks , how much of the time has your physical health or emotional problems interfered with your social activities (like visiting with friends, relatives, etc.)?
Survey E
19
All ofthe time
Most ofthe time
Some ofthe time
A little ofthe time
None ofthe time
a Have you felt calm and
peaceful?................................................... 1................. 2................. 3................. 4................. 5
b Did you have a lot of energy?................... 1................. 2................. 3................. 4................. 5
c Have you felt downhearted and depressed?.......................................... 1................. 2................. 3................. 4................. 5
All ofthe time
Most ofthe time
Some ofthe time
A little ofthe time
None ofthe time
1 2 3 4 5
TREATMENT GROUP ONLY
1. Were the instructions in the “perineal self-acupressure” pamphlet clear? Yes ⎕ No ⎕
2. Did you try using the “perineal self-acupressure” technique?Yes ⎕ No ⎕
If no, please indicate why: _______________________________________________________________
3. For the 4 weeks, please estimate the frequency with which you used the technique:
Times Per Week
0 1–2 3–4 5–7 >7
Perineal self-acupressure ⎕ ⎕ ⎕ ⎕ ⎕
4. If you used the technique, please answer the following questions.
a. Did you have difficulty using the technique?Yes ⎕ No ⎕
If yes, please indicate why: ________________________________________________________
b. Did the “perineal self-acupressure” technique help you defecate more easily?Yes ⎕ No ⎕
c. Did the “perineal self-acupressure” technique make defecating less painful? Yes ⎕ No ⎕
d. Did the “perineal self-acupressure” technique help you break up, soften, or pass your stools? Yes ⎕ No ⎕
e. Do you feel that the technique helped you to avoid having a hemorrhoid or lessened the impact of an existing hemorrhoid?
Yes ⎕ No ⎕
If no, please indicate why: _________________________________________________________
f. Do you feel that the technique helped you to avoid or better manage the effects of constipation?
Yes ⎕ No ⎕
20
If no, please indicate why: _________________________________________________________
g. Will you continue to perform “perineal self-acupressure” Yes ⎕ No ⎕
h. Will you recommend this method to family members/friends? Yes ⎕ No ⎕
5. If you didn’t use the technique, please tell us why:
__________________________________________________________________
6. Please provide any additional comments about your experience with the “perineal self-acupressure” pamphlet including suggestions on how we might make the instructions clearer.
________________________________________________________________________________________________________________________
21
Online Appendix 4
There are 9 sections to this appendix.
A. LOCF (all individual questions)B. Complete case analysisC. Multiple imputation analysisD. Seemingly unrelated regression analysis (with LOCF dataset)E. Ordered probit analysis (with LOCF dataset)F. Treatment group only questionnaire analysis G. Instrumental variables analysis (with LOCF dataset)H. Demographics interaction analysis (with LOCF dataset)I. Pre-treatment symptom interaction analysis (with LOCF dataset)
22
Section A: LOCF (all individual questions)
Table A shows the treatment effect for each individual question asked. These are difference-in-different regressions with individual fixed effects. These fixed effects effectually difference the pre- and post-treatment data. N=200.
Table A
23
Note: These are difference-in-difference regressions with individual fixed effects.
24
Section B: Complete case analysis
In this analysis, we kept only the respondents who were not lost to follow up (N=182). Table B below shows the treatment effects for that group only.
Table B
Note: These are difference-in-difference regressions with individual fixed effects.
Section C: Multiple imputation analysis
25
In this analysis, we use multiple imputation methodology to imputes the missing values from participants who were lost to follow-up. We utilized 100 imputed datasets to do so, using the commands under mi in Stata. Table C presents the treatment effect results.
Table C
Section D: Seemingly unrelated regressions
26
In this robustness analysis, we use seemingly unrelated regression specifications to account for correlated residuals in the models. The results (LOCF dataset) are presented in Table D. Note that the variable PAC all is not included, as it is collinear with the four PAC subscales.
Table D
Note: These are difference-in-difference regressions with individual fixed effects.
Section E: Ordered probit analysis
27
Because many of the questions that were asked that respondents pick discrete numbers or intensities of certain subjective measures (PAC-QOL and SF12 questionnaires), the ordered probit methodology is an appropriate alternative, as it takes the integer nature of the data into account. Table E presents the results of these regressions for individual question from the two questionnaires mentioned above. We do not include any indices, as these are continuous variables.
Table E
28
Note: These are difference-in-difference regressions with individual fixed effects.
Section F: Treatment group only questionnaire results
29
Table F shows the results for the treatment group only questionnaire, administered after the treatment. We have imputed data for the 4 uncompleted follow surveys for treatment group members lost to follow up.
Table F
Section G: Instrumental variables analysis
30
Instead of the intention-to-treat analysis conducted in the paper, here we present an instrumental variables approach, in which we use the treatment assignment as an instrument for the use of the technique. This latter variable appears in the reports made by respondents in the treatment-only survey. We assume that the control group does not use the technique. Because treatment group assignment is (1) random and unrelated to observables and (2) is correlated with technique use (see section 6), it is an appropriate instrument. Table 6 presents IV estimate for the indicator variable used technique which is equal to 1 if the patient indicated that they used the technique. This is the LOCF dataset. Participants with missing data were assumed to not use the technique.
First stage: coefficient on treatment is .88*** (t = 18.96), R2 = 0.7857.
Table G
Stage 1:
Stage 2:
Note: The d_ before the variable indicates that the pre-treatment observation was subtracted from the post- treatment observation.
Section H: Demographic interactions
31
Table H presents the basic diff-in-diff regression (N=200) for the LOCF dataset, with the treatment*post variable interacted with demographic dummies.
Table H
Note: These are difference-in-difference regressions with individual fixed effects.
Section I: Pre-treatment symptom interactions
32
Table I shows the results of interactions of a differenced dataset (N=100), with the pre-treatment variables as well as pre-treatment variables interacted with a treatment dummy as controls. Pre-treatment physical discomfort and hemorrhoid scores affect the change from pre- to post-treatment (P < .01).
Table I
Note: The d_ before the variable indicates that the pre-treatment observation was subtracted from the post- treatment observation.
Online Appendix 5
33
The Stata code below will recreate the numbers behind the tables in the paper and Online Appendix 4. Tables formatting has been done in Excel, and will not be represented in the output from the file below.
********************************************************************************* .do file for * EFFECT OF PERINEAL SELF-ACUPRESSURE ON CONSTIPATION: THE PSAC RANDOMIZED CLINICAL TRIAL
* Last updated 19 June 2014
* Input files: * per_6_6_14 -- Main Data File* sf_v12_raw -- SF-12 Data
********************************************************************************
set more off
********************************************************************************
* Power calculation
power twomeans 1.85 (2.15 2.25 2.35), power(0.8 0.9) sd(.67) graph* orsampsi 0 .5, sd1(.67) sd2(.67) power(.9)
////////////////////////////////////////////////////////////////////////////////////////////////////DEMOGRAPHIC VARIABLE CREATION //////////////////////////////////////////////////////////////////////////////////////////////////////////////
use per_6_6_14, clear
gen row = _nsort rowgen break = "."by row: replace break = substr(A,2,1) if row <19by row: replace break = substr(A,3,1) if row >18gen post = (break == "B")replace post = 1 if A == "100B"
gen treatment = (GroupAssignment == "Treatment")
gen male = (Gender == "Male")gen over_40 = (Age >= 40)gen Am_Indian = (RaceAmericanIndianorAlaska == "Yes")
34
gen Asian = (RaceAsian == "Yes")gen Black = (RaceBlackorAfricanAmerican == "Yes")gen White = (RaceWhite == "Yes")
gen Hispanic = (Ethnicity == "Hispanic/Latino")
gen missing = 0replace missing = 1 if A == "14A" | A == "14B"replace missing = 1 if A == "23A" | A == "23B"replace missing = 1 if A == "39A" | A == "39B"replace missing = 1 if A == "77A" | A == "77B"replace missing = 1 if A == "79A" | A == "79B"replace missing = 1 if A == "83A" | A == "83B"replace missing = 1 if A == "91A" | A == "91B"replace missing = 1 if A == "92A" | A == "92B"replace missing = 1 if A == "99A" | A == "99B"
sort missing
save full_vars_per_6_6_14, replace
///////////////////////////////////////////////////////////////////////////////////////////////////////////// CLEAN OUTCOME VARIABLES //////////////////////////////////////////////////////////////////////////////////////////////////////////
use full_vars_per_6_6_14, clear
//Paitient Assessment of Constipationreplace Question7 = "." if Question7 == "I eat things instead that help"replace Question21 = "." if Question21 == "Not Marked"
forvalues i = 1/28 {rename Question`i' Q_PAC_`i'}
//Bowel Function Indexrename AM Q_BFI_ease_of_defecation rename AN Q_BFI_feeling_incom_bowel_evac rename AO Q_BFI_constipation
//Hemorrhoid Scorerename Inlast4weeknewhemorrhoidse Q_BFI_Hemreplace Q_BFI_Hem = "1" if Q_BFI_Hem == "Yes"replace Q_BFI_Hem = "0" if Q_BFI_Hem == "No"rename bleeding Q_BFI_bleeingrename itching Q_BFI_itchingrename pain Q_BFI_pain
//SF-12
35
rename AT Q_HWB_1rename Question2a Q_HWB_2arename Question2b Q_HWB_2brename Question3a Q_HWB_3arename Question3b Q_HWB_3brename Question4a Q_HWB_4arename Question4b Q_HWB_4brename BA Q_HWB_5rename Question6a Q_HWB_6arename Question6b Q_HWB_6brename Question6c Q_HWB_6crename BE Q_HWB_7
//Treatment Group Instrumentrename BF Q_TGI_1rename BG Q_TGI_2rename BH Q_TGI_3rename BI Q_TGI_4arename BJ Q_TGI_4brename Question4c Q_TGI_4crename Question4d Q_TGI_4drename Question4e Q_TGI_4erename Question4f Q_TGI_4frename Question4g Q_TGI_4grename Question4h Q_TGI_4hrename BQ Q_TGI_5rename BR Q_TGI_6
foreach v of varlist Q_* {replace `v' = "" if `v' == "Missing 2nd Surveys"}
destring Q_*, replace
egen person = sieve(A), keep(numeric)destring person, replace
gen treatment_post = treatment*postgen tr_post_male = treatment_post * malegen tr_post_over_40 = treatment_post * over_40gen tr_post_Hispanic = treatment_post * Hispanicgen tr_post_Asian = treatment_post * Asiangen tr_post_Black = treatment_post * Blackgen tr_post_Am_Indian = treatment_post * Am_Indian
//Generate the aggregated scoresgen PAC_phys_discomfort = (Q_PAC_1 + Q_PAC_2 + Q_PAC_3 + Q_PAC_4) / 4gen PAC_psyc_discomfort = (Q_PAC_5 + Q_PAC_6 + Q_PAC_7 + Q_PAC_8 + Q_PAC_9 + Q_PAC_10 + Q_PAC_11 + Q_PAC_12) / 8
36
gen PAC_psyc_worries = (Q_PAC_13 + Q_PAC_14 + Q_PAC_15 + Q_PAC_16 + Q_PAC_17 + Q_PAC_18 + Q_PAC_19 + Q_PAC_20 + Q_PAC_21 + Q_PAC_22 + Q_PAC_23) / 11gen PAC_psyc_satisfaction = (Q_PAC_24 + Q_PAC_25 + Q_PAC_26 + Q_PAC_27 + Q_PAC_28) / 5gen PAC_all = (Q_PAC_1 + Q_PAC_2 + Q_PAC_3 + Q_PAC_4 + Q_PAC_5 + Q_PAC_6 + Q_PAC_7 + Q_PAC_8 + Q_PAC_9 + Q_PAC_10 + Q_PAC_11 + Q_PAC_12 + ///Q_PAC_13 + Q_PAC_14 + Q_PAC_15 + Q_PAC_16 + Q_PAC_17 + Q_PAC_18 + Q_PAC_19 + Q_PAC_20 + Q_PAC_21 + Q_PAC_22 + Q_PAC_23 + (20 - (Q_PAC_24 + Q_PAC_25 + Q_PAC_26 + Q_PAC_27 + Q_PAC_28))) / 28
gen bowel_function_score = (Q_BFI_ease_of_defecation + Q_BFI_feeling_incom_bowel_evac + Q_BFI_constipation) / 3gen hemmor_score = (Q_BFI_bleeing + Q_BFI_itching + Q_BFI_pain) / 3
save per_clean, replace
* Merge SF-12 Datause sf_v12_raw, clear
drop if _n == 1keep recordid pf_nbs rp_nbs bp_nbs gh_nbs vt_nbs sf_nbs re_nbs mh_nbs pcs mcsrename recordid A
merge m:m A using per_clean
save per_clean_2, replace
/////////////////////////////////////////////////////////////////////////////////////////////////// Creating LOCF and MI Datasets //////////////////////////////////////////////////////////////////////////////////////////////////////////////
* LOCF dataset
use per_clean_2, clearsort A
foreach v of varlist pf_nbs rp_nbs bp_nbs gh_nbs vt_nbs sf_nbs re_nbs mh_nbs pcs mcs ///Q_* bowel_function_score hemmor_score ///PAC_phys_discomfort PAC_psyc_discomfort PAC_psyc_worries PAC_psyc_satisfaction PAC_all {replace `v' = `v'[_n-1] if missing == 1 & post == 1}
save LOCF_data_full, replace
* Reshaped LOCF dataset
use LOCF_data_full, clear
37
set more off
sort A
keep PAC_p* PAC_all bowel_function_score hemmor_score Q_PAC_* Q_BFI_* Q_HWB_* person post treatment pcs mcs
foreach v of varlist Q_PAC_* Q_BFI_* Q_HWB_* {rename `v' `v'_}
reshape wide PAC_p* PAC_all bowel_function_score hemmor_score Q_PAC_* Q_BFI_* Q_HWB_* pcs mcs, i(person) j(post)
gen d_PAC_phys_discomfort = PAC_phys_discomfort1 - PAC_phys_discomfort0gen d_PAC_psyc_discomfort = PAC_psyc_discomfort1 - PAC_psyc_discomfort0gen d_PAC_psyc_worries = PAC_psyc_worries1 - PAC_psyc_worries0gen d_PAC_psyc_satisfaction = PAC_psyc_satisfaction1 - PAC_psyc_satisfaction0gen d_PAC_all = PAC_all1 - PAC_all0gen d_bowel_function_score = bowel_function_score1 - bowel_function_score0gen d_hemmor_score = hemmor_score1 - hemmor_score0gen d_pcs = pcs1 - pcs0gen d_mcs = mcs1 - mcs0
save reshaped_LOCF, replace
* MI dataset
use per_clean_2, clearset more off
sort A
keep PAC_p* PAC_all bowel_function_score hemmor_score Q_PAC_* Q_BFI_* Q_HWB_* person post treatment pcs mcs Age male Am_Indian Asian Black White Hispanicforeach v of varlist Q_PAC_* Q_BFI_* Q_HWB_* {rename `v' `v'_}reshape wide PAC_p* PAC_all bowel_function_score hemmor_score Q_PAC_* Q_BFI_* Q_HWB_* pcs mcs, i(person) j(post)
gen d_PAC_phys_discomfort = PAC_phys_discomfort1 - PAC_phys_discomfort0gen d_PAC_psyc_discomfort = PAC_psyc_discomfort1 - PAC_psyc_discomfort0gen d_PAC_psyc_worries = PAC_psyc_worries1 - PAC_psyc_worries0gen d_PAC_psyc_satisfaction = PAC_psyc_satisfaction1 - PAC_psyc_satisfaction0gen d_PAC_all = PAC_all1 - PAC_all0gen d_bowel_function_score = bowel_function_score1 - bowel_function_score0gen d_hemmor_score = hemmor_score1 - hemmor_score0gen d_pcs = pcs1 - pcs0
38
gen d_mcs = mcs1 - mcs0
save reshaped_for_mi, replace
use reshaped_for_mi, clearmi set flongmi register imputed d_PAC_phys_discomfort d_PAC_psyc_discomfort d_PAC_psyc_worries d_PAC_psyc_satisfaction d_bowel_function_score d_hemmor_score d_pcs d_mcsmi register regular treatment Age male Am_Indian Asian Black White Hispanic bowel_function_score0 hemmor_score0 PAC_phys_discomfort0 PAC_psyc_discomfort0 PAC_psyc_worries0 PAC_psyc_satisfaction0 PAC_all0 pcs0 mcs0mi describe
set seed 49230
mi impute mvn d_PAC_phys_discomfort d_PAC_psyc_discomfort d_PAC_psyc_worries d_PAC_psyc_satisfaction d_bowel_function_score d_hemmor_score d_pcs d_mcs /// = treatment Age male Am_Indian Asian Black White Hispanic bowel_function_score0 hemmor_score0 PAC_phys_discomfort0 PAC_psyc_discomfort0 PAC_psyc_worries0 PAC_psyc_satisfaction0 PAC_all0 pcs0 mcs0, add(100)
save MI_data_full, replace
////////////////////////////////////////////////////////////////////////////////////////////////////////////// Table 1 ///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
use LOCF_data_full, clear
keep if post == 0
sum male Age Am_Indian Asian Black Hispanic White
sum PAC_p* PAC_all bowel_function_score hemmor_score pcs mcs
local x 1set more offforeach v of varlist male Age Am_Indian Asian Black Hispanic White PAC_p* PAC_all bowel_function_score hemmor_score pcs mcs {ttest `v', by(treatment)putexcel A`x'=("`v'") B`x'=(r(mu_1)) C`x'=(r(mu_2)) D`x'=(r(p)) using tr_t_tests, modifylocal x = `x' + 1}
////////////////////////////////////////////////////////////////////////////////////////////////////////////// Table 2 /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
39
use reshaped_LOCF, clear
foreach v of varlist d_* {reg `v' treatment, robust}
////////////////////////////////////////////////////////////////////////////////////////////////////////////// eAppendix 5 /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
* Section A
use LOCF_data_full, clearset more off
foreach var of varlist PAC_p* PAC_all bowel_function_score hemmor_score pcs mcs *Q_PAC_* Q_BFI_* Q_HWB_* {reg `var' post treatment_post i.person, robustoutreg2 using per_1, excel keep(post treatment_post)}
* Section B
use per_clean_2, clearset more offdrop if missing == 1
foreach var of varlist PAC_p* PAC_all bowel_function_score hemmor_score pcs mcs *Q_PAC_* Q_BFI_* Q_HWB_* {reg `var' post treatment_post i.person, robustoutreg2 using per_2, excel keep(post treatment_post)}
* Section C
use MI_data_full, clear
replace d_PAC_all = ((4 * d_PAC_phys_discomfort + 8 * d_PAC_psyc_discomfort + 11 * d_PAC_psyc_worries + 5 * d_PAC_psyc_satisfaction) / 28) if d_PAC_all == .
foreach v of varlist d_* {mi estimate: reg `v' treatment, robust}
40
* Section D
use LOCF_data_full, clearset more off
set matsize 6000 sureg (PAC_phys_discomfort post treatment_post i.person) ///(PAC_psyc_discomfort post treatment_post i.person) ///(PAC_psyc_worries post treatment_post i.person) ///(PAC_psyc_satisfaction post treatment_post i.person) ///(bowel_function_score post treatment_post i.person) ///(hemmor_score post treatment_post i.person) ///(pcs post treatment_post i.person) ///(mcs post treatment_post i.person)
outreg2 using per_4, excel replace keep(post treatment_post)
* Section E
use LOCF_data_full, clearset more off
foreach var of varlist Q_PAC_* Q_HWB_* {oprobit `var' post treatment_post i.personoutreg2 using per_5, excel keep(post treatment_post)}
* Section F
use LOCF_data_full, clearkeep if post == 1keep if treatment == 1
tab Q_TGI_1gen instructions_clear = (Q_TGI_1 == "Yes")replace instructions_clear = 0 if Q_TGI_1 == ""
tab Q_TGI_2gen used_technique = (Q_TGI_2 == "Yes")tab used_technique
tab Q_TGI_3tab time_per_week_rangegen times_per_week_avg = 0replace times_per_week_avg = 1.5 if time_per_week_range == "1 to 2"
41
replace times_per_week_avg = 3.5 if time_per_week_range == "3 to 4"replace times_per_week_avg = 6 if time_per_week_range == "5 to 7"replace times_per_week_avg = 8 if time_per_week_range == ">7"
tab Q_TGI_4agen diff_using_tech = 1replace diff_using_tech = 0 if Q_TGI_4a == "No" | Q_TGI_4a == "" | Q_TGI_4a == "No (written)"
tab Q_TGI_4bgen def_more_easily = (Q_TGI_4b == "Yes" | Q_TGI_4b == "Yes (explanation)" | Q_TGI_4b == "Yes (written)")
tab Q_TGI_4c gen def_less_painful = (Q_TGI_4c == "Yes" | Q_TGI_4c == "Yes (explanation)")
tab Q_TGI_4dgen def_help_pass_stool = (Q_TGI_4d == "Yes" | Q_TGI_4d == "Yes (explanation)")
tab Q_TGI_4egen help_hemm = (Q_TGI_4e == "Yes" | Q_TGI_4e == "Yes (explanation)" | Q_TGI_4e == "Yes (written)")
tab Q_TGI_4fgen help_manage_const = (Q_TGI_4f == "Yes")
tab Q_TGI_4ggen continue_to_preform = (Q_TGI_4g == "Yes" | Q_TGI_4g == "Yes (explanation)" | Q_TGI_4g == "Yes (written)")
tab Q_TGI_4hgen recommend_fam_friends = (Q_TGI_4h == "Yes" | Q_TGI_4h == "Yes (written)")
sum instructions_clear used_technique times_per_week_avg ///diff_using_tech def_more_easily def_less_painful def_help_pass_stool help_hemm help_manage_const continue_to_preform recommend_fam_friends //if missing == 0
keep person post used_technique time_per_week_range times_per_week_avgsave treatment_only_post_data, replace
* Section G
use reshaped_LOCF, clearset more off
merge m:m person using treatment_only_post_data
replace used_technique = 0 if used_technique == .replace times_per_week_avg = 0 if times_per_week_avg == .
42
reg used_technique treatmentoutreg2 using per_7, replace excel
foreach v of varlist d_* {ivregress 2sls `v' (used_technique = treatment), robustoutreg2 using per_7, excel }
* Section H
use LOCF_data_full, clearset more off
foreach var of varlist PAC_all PAC_p* bowel_function_score hemmor_score pcs mcs {reg `var' post treatment_post tr_post_male tr_post_over_40 tr_post_Hispanic tr_post_Asian tr_post_Black tr_post_Am_Indian i.person, robustoutreg2 using per_8, excel keep(post treatment_post tr_post_male tr_post_over_40 tr_post_Hispanic tr_post_Asian tr_post_Black tr_post_Am_Indian)}
* Section G
use reshaped_LOCF, clearset more off
gen treatmentXPAC_phys_discomfort0 = treatment * PAC_phys_discomfort0gen treatmentXPAC_psyc_discomfort0 = treatment * PAC_psyc_discomfort0gen treatmentXPAC_psyc_worries0 = treatment * PAC_psyc_worries0gen treatmentXPAC_psyc_satisfaction0 = treatment * PAC_psyc_satisfaction0gen treatmentXbowel_function_score0 = treatment * bowel_function_score0gen treatmentXhemmor_score0 = treatment * hemmor_score0gen treatmentXpcs0 = treatment * pcs0gen treatmentXmcs0 = treatment * mcs0
foreach v of varlist d_PAC_phys_discomfort d_PAC_psyc_discomfort d_PAC_psyc_worries d_PAC_psyc_satisfaction d_bowel_function_score d_hemmor_score d_pcs d_mcs{reg `v' treatment PAC_phys_discomfort0 PAC_psyc_discomfort0 PAC_psyc_worries0 PAC_psyc_satisfaction0 bowel_function_score0 hemmor_score0 pcs0 mcs0 treatmentX*, robustoutreg2 using per_9, excel}
* END
43