SS B Reinstatement 0920-0969

Information Collection Request

Reinstatement

Monitoring Changes in Attitudes and Practices among Family Planning Providers and Clinics

OMB Number 0920-0969

Supporting Statement B

Contact Person:

Lauren B. Zapata, PhD, MSPH

Division of Reproductive Health

Centers for Disease Control and Prevention

4770 Buford Highway NE, MS F-74

Chamblee, GA 30341-3717

Phone: 770-488-6358

Fax: 770-488-6391

Email: [email protected]

September 26, 2018

Form 83-I

TABLE OF CONTENTS

B.

COLLECTION OF INFORMATION EMPLOYING STATISTICAL METHODS

B1. Respondent Universe and Sampling Methods

B2. Procedures for the Collection of Information

B3. Methods to Maximize Response Rates and Deal with Non response

B4. Test of Procedures or Methods to be Undertaken

B5. Individuals Consulted on Statistical Aspects and Individuals Collecting and/or Analyzing Data

REFERENCES

ATTACHMENTS

Attachment A-1 Section 301 of the Public Health Service Act (42 U.S.C. 241)

Attachment B-1 Summary of Survey Constructs

Attachment C-1 60-day FRN

Attachment C-2 Summary of Comments

Attachment D-1 Private-sector Initial Cover Letter

Attachment D-2 Public-sector Initial Cover Letter

Attachment E-1 2018-2019 Survey of Health Care Providers about Family Planning Attitudes and Practices (i.e., provider survey)

Attachment E-2 Screen Shots of Provider Survey

Attachment F-1 2018-2019 Survey for Administrators of Health Centers that Provide Family Planning (i.e., administrator survey)

Attachment F-2 Screen Shots of Administrator Survey

Attachment G-1 Private-sector Reminder Postcard for Provider Survey

Attachment G-2 Public-sector Reminder Postcard for Provider Survey

Attachment G-3 Public-sector Reminder Postcard for Administrator Survey

Attachment H-1 Email to Private-sector Physician Non-Respondents

Attachment I-1 Private-sector Follow-up Cover Letter

Attachment I-2 Public-sector Follow-up Cover Letter for Provider Survey

Attachment I-3 Public-sector Follow-up Cover Letter for Administrator Survey

Attachment J-1 Privacy Impact Assessment

Attachment J-2 IRB Non-research Determination

Attachment K-1 References

B. COLLECTIONS OF INFORMATION EMPLOYING STATISTICAL METHODS

B1. Respondent Universe and Sampling Methods

The proposed samples are intended to be nationally representative, as the sample frames will represent a complete census of the target populations.

• To sample private-sector physicians specializing in obstetrics and gynecology, family medicine, and adolescent medicine (specialties that provide the bulk of family planning services in the United States), we will use the AMA Physician Masterfile, which includes information on AMA member and nonmember board-certified physicians residing in the United States and United States territories.

• To sample public-sector family planning providers and health center administrators, we will use a database of publicly-funded family planning health centers, one that is regularly updated and maintained (i.e., Guttmacher Institute Database). As there is no database of health center administrators or health center providers, we will sample health centers; at each sampled health center, one provider and one administrator will be asked to complete surveys.

CDC and OPA will not build the list of sampled providers and health centers, but will instead purchase a list of sampled entities from an AMA database licensee (e.g., Medical Marketing Service, Inc) and the Guttmacher Institute, as was done in Phase 2 (OMB #0920-0969).

For each respondent type, the table below summarizes the sampling frame that will be used, the number of entities in the respondent universe, the desired number in the final sample (based on power calculations), the expected response rate, the number to be sampled (taking into account the desired number in the final sample and the expected response rate), and the sampling fraction to be used (taking into account the number to be sampled and the number of entities in the respondent universe).

Respondent Type		Sampling Frame	Number in Respondent Universe	Desired Number in Final Sample	Expected Response Rate a	Number to be Sampled	Sampling Fraction
(1) Private-sector office-based physicians
	OB/GYN physicians	AMA Physician Masterfile	34,426	500	50%	1,000	1/34
	Family medicine physicians	AMA Physician Masterfile	77,927	362	50%	725	1/107
	Adolescent medicine physicians	AMA Physician Masterfile	275	138	50%	275	1/1 b
TOTAL private-sector office-based physicians				1,000		2,000
(2) Public-sector health centers that provide family planning services
Title X clinics c		Guttmacher Institute database	4,095	1,000	50%	2,000	1/2
Non –Title X clinics c		Guttmacher Institute database	3,903	1,000	50%	2,000	1/2
TOTAL public-sector health centers that provide family planning services				2,000 (total of 4,000 responses: 2,000 providers and 2,000 administrators)		4,000** (total of 8,000: 4,000 providers and 4,000 administrators)

^a Response rates from our Phase 1 data collection (EPI AID No. 2010-024; OMB No. 0920-0008) were: OB/GYN physicians=52%, family medicine physicians=45%, adolescent medicine physicians=68%, and Title X clinic providers=77%. Response rates from our Phase 2 data collection (OMB Control Number 0920-0969) were: private-section physicians=40%, Title X clinic providers=66%, and non-Title X clinic providers=48%. Given that response rates fluctuate over time and slight changes in methodology between phases, we assume 50% response rate for all respondent types here.

^b For office-based adolescent medicine physicians, the entire universe will be sampled to ensure an adequate number of providers from this specialty being represented in the sample.

^c These consist of hospitals, health departments, Planned Parenthood clinics, community health centers, and ‘other’ clinic types; these strata of clinic types will be sampled proportionate to their representation in the universe.

**At each sampled public-sector health center, one provider and one administrator will be asked to complete surveys (i.e., 2,000 Title X clinic providers will be asked to complete the provider survey; 2,000 Title X clinic administrators will be asked to complete the administrator survey; 2,000 non-Title X clinic providers will be asked to complete the provider survey; and 2,000 non-Title X clinic administrators will be asked to complete the administrator survey).

For the two major respondent type categories (i.e., private-sector office-based physicians and public-sector health centers that provide family planning services [which will survey both providers and administrators on behalf of their health center), the estimates in the “desired number in final sample” column for the row “total” for that respondent type category (highlighted in table in bold), are derived from power calculations conducted when collecting our baseline data in Phase 1 (EPI AID No. 2010-024; OMB No. 0920-0008). We sought to power the sample to detect a 5% change from baseline levels based on a 2-tailed test at 5% significance with 80% power, assuming 50% non-response.

For private-sector office-based physicians, when drawing the sample for each specialty (i.e., OB/GYN, family medicine, adolescent medicine), the respondent universe will be sorted first by zip code and the sample fraction employed.

For public-sector health centers that provide family planning services, after sorting clinics by receipt of Title X funding status (i.e., Title X clinic versus non-Title X clinic), the universes will be sorted by clinic type (i.e., hospital, health department, Planned Parenthood clinic, community health center, and ‘other’), and then sorted by zip code (see below Figure). Clinic types will be sampled proportionate to their representation in the universe. For example, if 50% of Title X clinics are health departments, then health departments will represent 50% of the Title X sample. Again, at each sampled health center, one provider and one administrator will be asked to complete survey.

Database of Publicly-funded Family Planning Health Centers
Title X Clinics						Non-Title X Clinics
	Hospitals Health departments Planned Parenthood clinics Community health centers ‘Other’ clinic types						Hospitals Health departments Planned Parenthood clinics Community health centers ‘Other’ clinic types
(1) Provider Survey				(2) Administrator Survey		(1) Provider Survey				(2) Administrator Survey

B2. Procedures for the Collection of Information

Project staff at the CDC and OPA will obtain and provide to the data collection contractor the list of sampled private-sector office-based physicians and public-sector health centers that provide family planning services.

The data collection contractor will prepare the mailed survey packages and send to the 6,000 sampled private-sector physicians and public-sector health centers nationwide (recall that 1 survey will be sent to 2,000 private-sector physicians and 2 surveys [1 survey for clinic providers and 1 for clinic administrators] will be sent to 4,000 public-sector health centers).

The mailed survey packages will include a cover letter (Attachments D-1 and D-2) addressed personally to the physician or health center and will include a description of the assessment, address the importance of participation, and include a point of contact to direct inquiries.

For private-sector physicians, each mailed survey package will include a single survey (Attachment E-1) to be completed by the physician.

For public-sector health centers, each mailed survey package will include two surveys) – one to be completed by a clinician who provides family planning services to at least two women of reproductive age per week, on average (Attachment E-1), and the second to be completed by a health center administrator (Attachment F-1).

Each survey will contain a unique identification number (UID) printed on the first page of the survey; the UID will be assigned by the data collection contractor. CDC will not have access to any file linking names and addresses of physicians and health centers in our sample with their assigned UIDs. Each mailed survey will be accompanied by a postage-paid return envelope addressed to the contractor via a rented postal office box. Potential respondents will also be given the option to complete the surveys online (see Attachments E-2 and F-2 for screen shots).

Anticipating non-response, a reminder postcard will be sent to those who have not responded to the first mailing after approximately 2-4 weeks (Attachments G-1, G-2, and G-3). About one week later, the company overseeing management of the AMA Physician Masterfile will send an email reminder (Attachment H-1) to private-sector physician non-respondents. Similar outreach to non-respondents in public-sector health centers is not possible since email addresses are not included in the Guttmacher Institute database. A second copy of the survey, along with a follow-up cover letter (Attachments I-1, I-2, and I-3) will be sent to those who have not responded approximately 2-4 weeks after the reminder postcard. Phone calls will be made and emails sent (if email addresses are provided during phone call outreach) to those that have not responded to any of the contact attempts to encourage participation.

Data collected online will be downloaded into electronic databases (one for the provider survey and one for the health center administrator survey) on a regular basis. Paper-copy survey data will be entered into electronic databases (one for the provider survey and one for the health center administrator survey). The databases, stripped of any identifiers other than the UID, will be permanent federal records and will be maintained in accordance with CDC’s records control schedule (http://isp-v-maso-apps/RecSched/ViewSchedule.aspx?RID=29). Paper-copy surveys will be shredded within eight months after completion of data entry. Respondents will not be re-contacted after survey completion to validate any potentially unclear data elements.

The data collection contractor will develop data entry databases and analytic codebooks to facilitate valid and efficient data entry and analysis. For quality control purposes, 10-25% of paper-copy completed surveys will be double-entered.

The study design for this ICR is not experimental, but is instead a cross-sectional assessment.

B3. Methods to Maximize Response Rates and Deal with No Response

To maximize response rates, aspects of Dillman’s “Tailored Design Method” will be followed.

For example, each mailed survey package will include a cover letter that is addressed personally to the sampled physician or health center, and will include pieces of information that have been shown to be critical for enhancing response (e.g., the request, how the individual or health center was selected, and the usefulness of the information).

We will also request that select partner organizations (e.g., American College of Obstetricians and Gynecologists), if they choose, send a pre-survey email to their constituents. The content of the email will be left up to the discretion of the partner organization, but may include a positive and timely notice that the constituent may be receiving a request from the CDC and OPA to help with an important survey and encourage their participation.

Once initial survey packages are mailed to respondents, multiple contacts will be made to non-respondents by the data collection contractor to encourage participation. These include sending a reminder postcard approximately 2-4 weeks after the initial mailing, an email reminder to private-sector physicians approximately one week later, a second copy of the survey sent approximately 2-4 weeks after the reminder postcard, followed by phone calls and email contacts (if email addresses are provided during phone call outreach) to those that have not responded to any of the previous contact attempts.

Last, to maximize the response rate, as well as to provide important family planning information, a family planning provider tool will be sent to all sampled providers and clinics as part of the initial mailing and invitation to participate. The provider tool will not summarize information to be queried about in the surveys, so as to not bias the survey results. The provider tool may include one the following: recommended actions after late or missed combined hormonal contraception; a chart summarizing birth control method options; or infographic on making clinics teen-friendly.

As the proposed samples are intended to be nationally representative (see section B1), data will be weighted after data collection. The created weight variable will adjust for the probability of selection into the sample (given the varying sampling fractions to be employed), as well as nonresponse bias (given an anticipated response rate <80%). In addition, we will know basic information about each of the sampled physicians and health centers, which will be analyzed to understand general differences between respondents and non-respondents. For example, for the publicly-funded family planning health centers, we will know the type of health center (e.g., Planned Parenthood, hospital, community health center). For physicians, we will know background information such as specialty, gender, and graduation year from medical school.

B4. Tests of Procedures or Methods to be Undertaken

The data collection procedures and instruments are mostly the same as those employed during our US MEC baseline data collection in Phase 1 (EPI AID No. 2010-024; OMB No. 0920-0008) and during our Phase 2 assessment (OMB Control No. 0920-0969).

The data collection instruments were developed using input from internal and external consultants to improve clarity and validity. The instruments were then pilot tested with a sample of no more than nine individuals.

B5. Individuals Consulted on Statistical Aspects and Individuals Collecting and/or Analyzing Data

Name	Organization	Role
Brian Morrow*, (770) 488-6349, [email protected]	CDC/DRH	Statistician
Gary Jeng*, (770) 488-6216, [email protected]	CDC/DRH	Statistician
Michele Mandel*, (770) 488-6361, [email protected]	CDC/DRH	Data analyst
Cindy Reilly, Project Manager	Insignia Federal Group	Data collection
Bea Snidow, Data Analyst/Manger	Insignia Federal Group	Data collection
David Burns, Data Analyst/Manager	Insignia Federal Group	Data collection
Lauren Zapata*, (770) 488-6358, [email protected]	CDC/DRH	Data analyst
Maura Whiteman*, (770) 488-6293, [email protected]	CDC/DRH	Data analyst
Kate Curtis*, (770) 488-6387, [email protected]	CDC/DRH	Data analyst
Naomi Tepper*, (770) 488-6506, [email protected]	CDC/DRH	Data analyst
Tara Jatlaoui*, (770) 488-6537, [email protected]	CDC/DRH	Data analyst
Isabel Morgan*, (404) 498-1266, [email protected]	CDC/DRH	Data analyst

*Contributed to design of the data collection and analysis.

11