The next session of the annual HOS survey vendor training will take place in Spring 2021. To administer the HOS, conditionally-approved survey vendors must participate in the annual survey vendor training and successfully complete the post-training evaluation. Survey vendors receive an overview of survey vendor requirements including data submission and coding, quality oversight, and data management. Prior survey vendor training slides can be found on the Survey Administration page on the HOS website. Additional resources to learn more about the HOS such as technical reports, webinars, and manuals can be found on the Resources page.

The fielding phase of the 2021 HOS will begin in the summer after survey vendor training is complete. The 2021 HOS Approved Survey Vendors list will be found along with other timeline information on the HOS website Program page in the Survey Vendors section. The 2021 Survey Administration Memos will be found on the Survey Administration page. Links to copies of the HEDIS^® MY 2020, Volume 6: Specifications for the Medicare Health Outcomes Survey and the 2021 Quality Assurance Guidelines and Technical Specifications V2.5 will also be provided when available.

The highlighted sections of the table above depict the relationship between recent data collection and reporting periods. The 2021 Medicare Star Ratings data are from the Cohort 20 HOS measures and the HEDIS Effectiveness of Care measures that were collected in 2019. The 2021 Medicare Star Ratings will be used to set the 2022 quality bonus payments as shown in the green highlighted sections of the table. The yellow highlighted sections show how the 2022 Medicare Star Ratings will be used to set the 2023 quality bonus payments.

The most recently available HOS reports include:

Registered Health Plan Management System (HPMS) users have access to these reports through the HPMS site. The following CMS site has information about how to establish access to HPMS: www.cms.gov/Research-Statistics-Data-and-Systems/Computer-Data-and-Systems/HPMS/Overview. If you require assistance accessing reports in HPMS, please contact CMS via email at hpms_access@cms.hhs.gov.

There is a table available on the HOS website Survey Results page that provides general survey status information. The table includes cohort sample sizes and response rates for the baseline and follow up surveys that have been administered and reported through 2019.

The HOS website also provides downloadable sample HOS-M, Baseline, and Performance Measurement Reports with actual aggregate national benchmark data compared to example plan data. These sample reports may be downloaded from the Overview section on the Resources page.

Contact the HOS Team at hos@hsag.com to request data for your MAO or Program of All-Inclusive Care for the Elderly (PACE) organization. Encrypted data are distributed to designated recipients via secure file transfer.

Medicare Star Ratings are in part derived from the two functional health outcome measures and three effectiveness of care measures calculated from beneficiary responses to HOS questions. CMS uses these Star Ratings to help consumers compare health plans based on quality and performance and to reward top-performing health plans. The measure scores obtained through the HOS contribute to individual Star Ratings that are created on a five-star rating scale. When combined with other measurement results, these measure scores contribute to the summary-level Part C Star Ratings which are reported on the Medicare Plan Finder page (www.medicare.gov/plan-compare). Consumers can search for health plans in their geographic area and compare costs, coverage, and quality information.

The Modified Regression Estimate (MRE) is used for missing data imputation and scoring of the unadjusted scores. The algorithm uses complete cases (no items missing) to estimate a regression equation for cases with missing items, where only those items that are present are used. The final adjusted measures are based on the case-mix adjusted PCS and MCS change scores between baseline and follow up surveys, as well as death status. Please refer to Appendix 1 of the Sample PM Report under “Calculating Performance Measurement Results” located on the HOS website Survey Results page, which describes the questions included in the calculation of PCS and MCS scores, and the case-mix adjusted outcomes for the PM results. The PM results were computed using rigorous case mix/risk adjustment procedures. Case-mix adjustments are used so that all MAOs are as comparable as possible in terms of the socio-demographic characteristics (age, gender, race, etc.), chronic medical conditions, baseline health status, and other design variables obtained from the survey. The coefficient tables of the covariates used in the series of case-mix models for expected death, PCS, and MCS (“Medicare HOS Performance Measurement Coefficient Tables”) are available on the HOS website Survey Results page.

Further details about the HOS variables (e.g., race and ethnicity) are provided in the Performance Measurement DUG that is provided to MAOs with their requested data or refer to the online document.

For more information and for scoring algorithms, go to the VR-12 developer’s website at https://www.bu.edu/sph/about/departments/health-law-policy-and-management/research/vr-36-vr-12-and-vr-6d/.

The Surveillance, Epidemiology, and End Results-MHOS (SEER-MHOS) is a linked data set created in collaboration with CMS and the National Cancer Institute (NCI). The SEER-MHOS data set links the SEER cancer registries with MAO enrollees who participated in the HOS in order to provide detailed clinical, demographic, and cause of death information for Medicare beneficiaries with cancer. Additionally, Medicare claims during the beneficiary’s Medicare eligibility (until death) are included in the data. The linked SEER-MHOS data resource includes HOS data covering the years from 1998 through 2018. For the past several years, NCI has made SEER*Stat software available that identifies SEER patients who responded to the HOS as well as the number of surveys they completed before and after cancer diagnosis. In 2021, NCI is developing a web-based sample size estimator that will allow investigators to quickly estimate research sample sizes as they consider using the SEER-MHOS linked data set. More information about the data set and how to apply for a data use agreement (DUA) can be found on the SEER-MHOS website. The following research articles highlighted in this section use the SEER-MHOS as their data source.

Chronic conditions, including various cancers, may impact HRQOL and require consistent access to resources for care. This makes location an important factor. A new article that used SEER-MHOS data, “Rural–urban differences in health-related quality of life: patterns for cancer survivors compared to other older adults” was released in November 2020 and may be found here.[1]

[1] Moss, J. L., Pinto, C. N., Mama, S. K., et al. Rural–urban differences in health-related quality of life: patterns for cancer survivors compared to other older adults. Quality of Life Research. 2020: 1-13. Available at: https://link.springer.com/article/10.1007/s11136-020-02683-3. Accessed on January 12, 2021.

Considering the current climate in healthcare, HRQOL is becoming increasingly important to overall health. HRQOL may also be used as an indicator to direct health-related efforts, which is crucial for targeting particular diseases and chronic conditions such as cancer. A new article for a case-control study that used the SEER-MHOS data, “Health-related quality of life and medical comorbidities in older patients with pancreatic adenocarcinoma: An analysis using the 1998–2011 Surveillance, Epidemiology, and End Results-Medicare Health Outcomes Survey data” was released in May 2020 and may be found here.[2]

[2] Wong, S. S., Hsu, F. C., Avis, N. E., et al. Health-related quality of life and medical comorbidities in older patients with pancreatic adenocarcinoma: An analysis using the 1998–2011 Surveillance, Epidemiology, and End Results-Medicare Health Outcomes Survey data. Journal of geriatric oncology. 2020: 11(4), 633-639. Available at: https://www.sciencedirect.com/science/article/abs/pii/S1879406819301201. Accessed on Jan 12, 2021.

The 2021 survey administration documents are available via the Survey Administration link. The documents include the HOS and HOS-M survey administration memos for MAOs and PACE organizations, and the HOS exclusion memo. Information about HOS program requirements may be found in the HEDIS MY 2020, Volume 6: Specifications for the Medicare Health Outcomes Survey Manual and 2021 Quality Assurance Guidelines and Technical Specifications Manual, Version 2.4, available on the Resources page. Information about HOS survey vendors is available on the Program page in the Survey Vendors section. Survey vendors approved to administer the 2021 HOS survey will be posted during the spring of 2021.

*Tentative year for data distribution, based on current timeline projections and CMS guidelines.

General information about the available HOS data files, variables contained in each file, and requirements for obtaining files may be found on the Research Data Files page of the HOS website. There are three categories of research data files, and they are described in more detail below.

HOS Public Use File (PUF):

Limited Data Set (LDS) File:

Research Identifiable File (RIF):

All MAOs required to report HOS will have the option of oversampling for the 2021 Cohort 24 Baseline survey. Oversampling involves surveying a sample of members that is larger than the required sample size of 1,200. There is currently no upper limit on oversampling for HOS; however, this guidance may change in future survey administration years. Oversampling can only occur at the contract level for the Baseline sample. The follow-up sample is still selected based on follow-up members with eligible PCS and MCS scores from two years prior. Oversampling requests should be expressed as a whole number percentage of the sample size. MAOs must notify the HOS Project Team at NCQA (hos@ncqa.org) of oversampling requests by the due date that is specified in the HOS documentation memo, and all oversampling requests are subject to approval by CMS.

The definition of a completed survey for the HOS-M Feedback Report is based on the VR-12 summary measures, while a completed survey for the frailty assessment is based on the ADL questions. Therefore, the ADL distributions derived for each report may differ somewhat due to the differing sample selection. Questions about the HOS-M Feedback reports may be directed to the hos@hsag.com mailbox, and questions about the Risk Adjustment reports may be directed to the CMS mailbox at RiskAdjustmentPolicy@cms.hhs.gov.

The annual Medicare HOS data collection generally occurs across a designated four-month period. Health plans should not field other HOS-like surveys during, and four weeks prior to, the HOS data collection period, except for other CMS sponsored surveys (i.e., CAHPS). This is the time frame that is sometimes referenced as the “blackout period.” Specific dates for the annual data collection period can be found on the Program page of the HOS website in the Program Timeline section.