OMB Supporting Statement(B)

Supporting Statement

Section B

NATIONAL CENTER FOR COMPLEMENTARY AND ALTERNATIVE
MEDICINE

NCCAM CUSTOMER SERVICE DATA COLLECTION

Request for Revision

August 27, 2009

Table of Contents

Introduction 3

B.1.Respondent Universe and Sampling Methods 3

B.2.Information Collection Procedures/Limitations of the Study 4

B.2.1.Statistical Methodology for Stratification and Sample Selection 5

B.2.2.Estimation Procedure 5

B.2.3.Degree of Accuracy Needed for the Purpose Described in the Justification 6

B.2.4.Unusual Problems Requiring Specialized Sampling Procedures 6

B.2.5.Use of Periodic (Less Frequent Than Annual) Data Collection Cycles 6

B.3.Methods for Maximizing the Response Rate and Addressing Issues of Nonresponse 6

B.4.Tests of Procedures or Methods 6

B.5.Names and Telephone Numbers of Individuals Consulted 7

Introduction

For the telephone survey, a simple random sampling technique is utilized to select a representative sample of callers to the NCCAM telephone information service. The data collected from the sampled respondents are used to characterize callers and ascertain their satisfaction with NCCAM materials and information. The use of a sample minimizes respondent burden by asking the survey questions of a sampling of callers rather than interviewing a census of all callers. The population of callers and the sampling plan are described below.

B.1.Respondent Universe and Sampling Methods

Telephone Survey. The population for the telephone survey consists of callers to the NCCAM telephone information service. Based on summary statistics for September 2003 to December 2008, an average of 648 callers per month telephoned the NCCAM telephone information service. The projected annual number of callers is 7,771.

NCCAM is interested in continuing to ascertain how callers learned of the NCCAM telephone information service and how satisfied callers are with the information provided. In order to meet the needs of the diverse population of callers, NCCAM would like to assess both of these topics by distinct user types (user types of interest are listed in Table B.1-1). These user types are expected to have different needs and expectations regarding information obtained from NCCAM.

Estimated annual population counts for these user types, based on survey results to date, are shown in Table B.1-1. It is estimated that nearly 9 of every 10 callers (88 percent) fall into one of three groups: patient, spouse/relative/friend of patient, and general public. Other callers, such as CAM practitioners, other health care providers, and researchers, acquire information for professional use.

B.1-1 Population of Callers to NCCAM Telephone Information Service
User type	Annual estimated callers
	Number	Percentage
Patient	3,921	51
Spouse/relative/friend of patient	1,491	19
General public	1,401	18
CAM practitioner	158	2
Other health care provider	345	4
Researcher or grant applicant	123	2
Journalist/media professional	9	0.1
Student	188	2
Other	134	2
TOTAL	7,771	100

An approximate 25 percent sampling rate (actually 23 percent) of callers is utilized, yielding a starting sample size of 1,787 callers. Assuming a gross response rate of 55 percent based on survey results to date, the actual number of callers participating in the interview is estimated to be 1,983.

B.2.Information Collection Procedures/Limitations of the Study

Telephone Survey. Approximately one-quarter of all callers are asked to participate in a short interview after receiving information from an Information Specialist. The callers are selected with a random number generator. A pop-up window appears on the Information Specialist's computer screen, indicating that the caller has been selected for an interview. Callers are asked to respond to each question on the survey instrument, except for the question on education level. Only patients, spouses/relatives/friends of patients, and the general public are queried as to their highest level of education. The estimated distributions of the population and the sample of completed interviews are shown in Table B.2-1.

A total sample size of 983 callers provides a precision level for an estimated overall proportion of respondents selecting a particular answer that is no greater than 3 percent at the 95 percent confidence level. Precision levels for specific user types vary from 4 to 7 percent for higher-volume user types such as patients, spouses/relatives/friends of patients, and the general public to 14 percent for lower-volume callers such as other health care providers.

B.2–1 Sample Distribution of Callers to NCCAM Telephone Information Service by User Type
User type	Estimated annual callers
		Number asked to participate in interview	Number expected to complete interview
	Number
Patient	3,921	902	496
Spouse/relative/friend of patient	1,491	343	189
General public	1,401	322	177
CAM practitioner	158	36	20
Other health care provider	345	79	44
Researcher or grant applicant	123	28	16
Journalist/media professional	9	2	1
Student	188	43	24
Other	134	31	17
TOTAL	7,771	1,787	983

B.2.1.Statistical Methodology for Stratification and Sample Selection

Phone survey participants are selected by a random number generator that results in

Information Specialists asking one of every four callers to participate.

B.2.2.Estimation Procedure

Due to the nature of the sampling and data collection procedures described in Sections

B.1 and B.2, estimates will take the form of simple frequency tables of respondent counts for each survey item.

B.2.3.Degree of Accuracy Needed for the Purpose Described in the Justification

The precision levels described earlier in section B.2 provide sufficient accuracy to assess customer satisfaction with NCCAM's phone service.

B.2.4.Unusual Problems Requiring Specialized Sampling Procedures

There are no unusual problems requiring specialized sampling procedures.

B.2.5.Use of Periodic (Less Frequent Than Annual) Data Collection Cycles

Data collection will not be less frequent than annual.

B.3.Methods for Maximizing the Response Rate and Addressing Issues of Nonresponse

Telephone Survey. The gross response rate is expected to be 55 percent, and the adjusted response rate is expected to be 70 percent, based on survey findings to date. A net response rate of approximately 70 percent should provide reliable estimates for a survey of this nature, particularly given the relatively large starting sample size. Followup of nonrespondents is not required.

B.4.Tests of Procedures or Methods

Each survey instrument was pretested with nine contractor staff to estimate response time and to refine the wording and design of the survey. Subsequent experience with administering the telephone survey to callers has verified the estimated response time.

B.5.Names and Telephone Numbers of Individuals Consulted

Contract personnel at Lockheed Martin implemented the sample design. These personnel train Information Specialists, conduct data collection and analysis, and develop statistical reports. NCCAM staff members provide direction and review contractor functions. Names, telephone numbers, and e-mail addresses of individuals consulted on statistical aspects of the design of this data collection include the following:

Nancy Hassett

Scientific Systems Analyst Staff

Lockheed Martin

Phone: 301-519-6372

E-mail: [email protected]

NCCAM Clearinghouse staff served as interviewers during the pretests.