Summary/Response to Public Comment

CMS Response to Public Comments Received for CMS-2017-0048

The Centers for Medicare & Medicaid Services (CMS) received several comments related to CMS-2017-0048 from one individual representing a health plan. This is the reconciliation of those comments.

Comment:

The commenter recommended CMS explore other ways of collecting the HOS data that would “provide real time data” and “also decrease the burden on the beneficiaries,” such as by adding the HEDIS-HOS items to the CAHPS or using the HRA.

Response:

CMS shares commenter’s concern about burdening our beneficiaries. However, adding the HEDIS-HOS items to the MAO-reported Health Risk Assessment (HRA) would defeat the HOS’s purpose as an unbiased patient-reported measure. While adding the HEDIS-HOS measures to the CAHPS survey would move up the data collection timeline by one month (CAHPS data are collected March through June; HOS data are collected April through July), the commenter is incorrect in assuming the HEDIS results could be distributed with the CAHPS results. Results of the four (4) HEDIS-HOS Effectiveness of Care measures (Management of Urinary Incontinence in Older Adults, Physical Activity in Older Adults, Fall Risk Management, and Osteoporosis Testing in Older Women) that are calculated by NCQA are not due to CMS until spring of the year following data collection and are distributed annually in the HOS Baseline Report May.

Action Taken: None.

Comment: The commenter recommended CMS reduce the “time lag” from HOS survey execution to final results.

Response:

The commenter pointed out the three year “time lag” between HOS baseline data collection and health plans receiving member-level results. Contributing to the perceived “lag” is the longitudinal component of the HOS; beneficiaries who complete the baseline HOS must be resurveyed two years later to generate HOS “outcome” measures. While health plans do wait three years for their performance measurement reports and beneficiary-level data, results of the baseline survey are distributed in May of the year following data collection, approximately nine months after data collection ends. Providing member-level data any earlier is not possible because doing so would prematurely reveal the sample.

Action Taken: None. CMS disagrees with the commenter’s statement that the delay in distributing member-level data does not allow quality improvement in a meaningful timeframe. There has been significant improvement in HOS scores in the 7 years that HOS has been included in the Star Ratings, indicating that as a group, health plans are acting on the data and achieving results.

Comment:

The commenter noted the HOS’s recent struggle to research targeted national response rates and recommended beneficiaries be given the option to complete the survey online as a means of potentially increasing response rates.

Response:

CMS has received requests for an online administration of the HOS and continues to explore a web data collection mode for the HOS and other surveys. To date, the mode remains problematic on several levels, including response rates which have been unacceptably low (lower even than mail only mode).

Action Taken: CMS will continue to explore a web data collection mode and looks forward to reviewing total survey error and other statistically significant markers of quality (i.e., item missingness, evaluation of response propensity by mode and patient characteristics, etc.) from other projects. In the event these web results prove satisfactory, CMS will explore additional uses of an online tool, potentially to include HOS.

Comment:

The commenter recommended the large plan exclusion not be eliminated from the HOS sample protocol in 2018 due to the presumed negative impact on response rates if a beneficiary were to be selected for the baseline survey in consecutive years.

Response:

CMS is mindful of beneficiary burden. The current HOS protocol states that in large plans (those with more than 3,000 eligible beneficiaries), beneficiaries are excluded from the baseline survey sample frame if they completed a survey in the previous year. The rationale for this decision, which made was more than 12 years ago, was ostensibly to reduce response burden so that the same beneficiaries were not asked to do a survey year after year. This decision, however, created a sample bias by excluding beneficiaries who legitimately should be part of the baseline sample frame. Moreover, it selectively excludes beneficiaries only in the largest plans, not the smaller ones. Ironically, it also eliminates the beneficiaries who are most likely to participate in the survey since they responded in the past.

The probability that any beneficiary in a large plan will be selected for the HOS survey is at most 40% (1,200/3,000) and declines to 2.3% for the average large plan with more than 52,000 members and 0.11% for the largest large plans with more than 1,100,000 members. Eliminating the large plan exclusion will eliminate the sample bias it created. Any impact on response rates is expected to be small but positive.

Action Taken: None. CMS plans to eliminate the large plan exclusion in order to eliminate the unintentional sample bias the exclusion created.

Comment:

The commenter noted that the HOS references multiple experience periods which may be confusing to beneficiaries and recommended that CMS implement a consistent timeframe throughout the instrument.

Response:

The commenter correctly noted the HOS references multiple experience periods. The HOS 3.0 Medicare questionnaire consists of several distinct components: the Veterans RAND 12-Item Health Survey (VR-12); four HEDIS® Effectiveness of Care measures; the Patient Health Questionnaire-2 (PHQ-2); four Healthy Days items (CDC HRQOL-4); clinical measures; chronic condition and demographic questions for case-mix and risk-adjustment; expanded measures of race, ethnicity, sex, primary language, and disability status required under Section 4302 of the Affordable Care Act; and additional self-reported health indicators. All of the items have specific purposes, and while timeframes do vary, each measure has undergone statistical testing and been validated for use with specific populations. In addition, CMS has previously conducted cognitive testing of the HOS 3.0 questionnaire and found no major issues.

Action Taken: None. CMS remains committed to ensuring respondents interpret the HOS instrument as intended and will contract for additional cognitive testing when revising and updating the HOS and as otherwise appropriate.

Comment:

The commenter recommended that CMS accommodate life changing events (such as loss of a loved one, admission to nursing home facility, or negative health diagnosis) in HOS survey analysis and results.

Response:

CMS holds a public comment period every year to solicit public comments on proposed quality measures. The public comment period provides an opportunity for the widest array of interested parties to provide input on the measures under development and can provide critical suggestions not previously considered. The public is encouraged to submit general comments and comments relevant to specific measures. At the end of the public comment period, all public comments are posted on the website along with a public comment summary report. The public comment dates for 2017 have passed, but the commenter is invited to submit suggestions during the 2018 comment period.

Action Taken: None. The commenter is invited to submit suggestions during the 2018 comment period (https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/MMS/CallforPublicComment.html).

Comment:

The commenter cited regional differences, such as the availability of good transportation, weather, overall health of the State, and the availability of community resources, that may advantage some areas when it comes to quality measures. The commenter also noted the HOS is available only in English, Spanish, and Chinese, which affects response rates for plans that serve beneficiaries who do not speak these three languages and disadvantages health plans that serve diverse populations. The commenter recommended CMS account for regional differences to ensure plans are not disadvantaged based on the populations served.

Response:

CMS already adjusts HOS data to control for many beneficiaries characteristics that are not under the control of the plan. Case-mix variables include age, gender, race, ethnicity, income, education, marital status, Medicaid status, SSI eligibility, homeowner status, chronic conditions, and baseline health status. Controlling for regional differences and infrastructure disparities is beyond the scope of the HOS.

The HOS is presently administered in three languages (English, Spanish, and Chinese) and also includes a two-part question that asks respondents about their primary spoken language. Respondents who answer “Other” are asked to answer by writing in (or over the telephone) the primary language spoken at home. CMS is in process of testing a Russian translation of the HOS. Other top responses to the language question include Tagalog, Vietnamese, and Korean, which may be slated for future translation and testing efforts.

Action: None. However, CMS and its contractors will continue to encourage health plans to partner with their members to maintain or improve their health status within the context of their life and regional circumstances.

Comment:

The commenter noted that some HOS questions center on conversations with a doctor or other provider, making the plan responsible for conversation at the provider’s office. The commenter recommends that CMS remove the burden on the health plan to manage the office visit between the beneficiary and the physician.

Response:

CMS feels that the health plan plays a significant role in patient-provider interactions and that effective interactions among providers and health systems may be crucial to improving outcomes. It is therefore important to assess the nature of interactions between patients and providers and health plan along multiple dimensions, including patient-provider communication. 

Action: None.

Comment:

The commenter recommended two changes to the revised HEDIS “Reducing the Risk of Falling” measure.

.

Response:

NCQA is the measure steward for the HEDIS-HOS measures, including Reducing Risk of Falling. NCQA reevaluated the measure and proposed changes to the measure that included updating the example interventions in the questions to align with the most current U.S. Preventive Services Task Force (USPSTF) guidelines. NCQA held invited public comments on the proposed changes in 2014 and 2015. CMS also solicited comments on the proposed changes in our “2015 Request for Comments.”

Action: None. The revision has been finalized by NCQA.

4