Vol.
-
No.
- -
2014
Journal of Pain and Symptom Management 1
Original Article
Hospices’ Use of Electronic Medical Records for Quality Assessment and Performance Improvement Programs Nan Tracy Zheng, PhD, Franziska S. Rokoske, PT, MS, M. Alexis Kirk, BA, Brieanne Lyda-McDonald, MSPH, and Shulamit L. Bernard, PhD Division of Health Services and Social Policy Research (N.T.Z.), RTI International, Waltham, Massachusetts; and Division of Health Services and Social Policy Research (F.S.R., M.A.K., B.L.-M., S.L.B.), RTI International, Research Triangle Park, North Carolina, USA
Abstract Context. Electronic medical records (EMRs) are increasingly viewed as essential tools for quality assurance and improvement in many care settings, but little is known about the use of EMRs by hospices in their quality assessment and performance improvement (QAPI) programs. Objectives. To examine the data sources hospices use to create quality indicators (QIs) used in their QAPI programs and to examine the domains of EMR-based QIs. Methods. We used self-reported QIs (description, numerator, and denominator) from 911 hospices nationwide that participated in the Centers for Medicare & Medicaid Services nationwide hospice voluntary reporting period. The data reflected QIs that hospices used for their internal QAPI programs between October 1 and December 31, 2011. We used the primary data sources for QIs reported by hospices and analyzed EMR-based QIs in terms of the quality domains and themes addressed. Results. EMRs were the most frequent data source for the QIs reported, followed by family survey and paper medical record. Physical symptom management was the largest quality domaindincluded in 51.5% of the reported EMR-based QIsdfollowed by patient safety and structure and process of care. Conclusion. Most participating hospices use EMRs for retrieving items needed for QI calculations. EMR-based QIs address various quality domains and themes. Our findings present opportunities for potential future reporting of EMR-based quality data. J Pain Symptom Manage 2014;-:-e-. Ó 2014 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Electronic medical record, electronic health record, hospice quality indicators, quality reporting
Address correspondence to: Nan Tracy Zheng, PhD, Aging, Disability and Long Term Care, Division of Health Services and Social Policy Research, RTI Ó 2014 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved.
International, 1440 Main Street, Suite 310, Waltham, MA 02451-1623, USA. E-mail: ntzheng@ rti.org 0885-3924/$ - see front matter http://dx.doi.org/10.1016/j.jpainsymman.2013.11.010
2
Zheng et al.
Introduction Medicare hospice services include palliation, comfort, and support services for Medicare beneficiaries with a life expectancy of six months or less and their families. In 2009, about 1.1 million beneficiaries (more than 40% of Medicare decedents in that year) received hospice services.1 Medicare hospice spending more than tripled between 2000 ($2.9 billion) and 2007 ($10 billion) and increased to $12 billion in 2009. At the same time, the number of hospice providers that are certified by the Centers for Medicare & Medicaid Services (CMS) has increased rapidly, particularly for-profit hospices, which increased by 150% between 2000 and 2010.1 Despite the high hospice utilization and the proliferation of hospice providers, the public and policymakers know relatively little about the quality of care provided by hospices. Developing, implementing, and maintaining an effective, ongoing, hospice-wide, data-driven quality assessment and performance improvement (QAPI) program has been a requirement in the Conditions of Participation since 2008, but whether hospices implement QAPI programs, and how many and what type of quality indicators (QIs) are included in the QAPI programs, are largely unknown. The Affordable Care Act required the U.S. Department of Health and Human Services to establish a Hospice Quality Reporting Program.2 CMS initiated a voluntary reporting period beginning in January 2012 for hospices nationwide to report the QIs that they used in their QAPI programs between October 1 and December 31, 2011.3 The purpose of this voluntary reporting period was to help CMS design the data collection approach for mandatory quality reporting starting in 2013. The data reported by hospices during the voluntary reporting period also provided the first opportunity to identify the QIs that they are using to meet QAPI requirements and the data sources for constructing their QIs. Electronic medical records (EMRs) are increasingly viewed as essential tools for quality assurance and improvement in a variety of care settings.4 The 2007 National Home Health and Hospice Care Survey found that approximately 49% of hospice agencies and 63% of agencies that provide both home health and hospice
Vol.
-
No.
- -
2014
care were using EMRs.5 The most common EMR functionalities or components that providers use are for recording patient demographics and clinical notes.6 Only one study investigated the use of EMRs specifically for the purposes of quality measurement or improvement in hospices. Cagle et al.7 reported on the use of electronic data (including EMRs) for QAPI purposes. Their survey of more than 650 hospice providers showed that larger, home-health agency-based, nonprofit, mixed urban/rural hospices had the highest rates of use of electronic data for quality measurement or improvement. Although this survey provided evidence that hospices were using electronic data as part of their quality measurement efforts, it did not collect data on what QIs hospices created using electronic data. The purposes of this study were to examine the data sources that hospices use to create individual QIs and to examine the domains of patient care addressed by EMR-based QIs.
Methods Data and Sample In late 2011, all hospices nationwide were made aware of their eligibility to participate in the voluntary reporting period through the CMS Hospice Wage Index Final Rule for FY2012 (CMS-1355-F, August 2011), CMS Open Door Forums, information posted on the CMS Hospice Quality Reporting Program Web site, and provider association newsblasts. A Web-based data collection form was created for hospices to submit data during the voluntary reporting period. Hospices were asked to enter details about the QIs they used if they had at least one patient care-related QI in their QAPI program in Quarter 4 of 2011 (October 1eDecember 31). For each QI, hospices provided the following information by either selecting from a drop-down menu or entering free text: patient care topic (e.g., pain management, psychosocial assessment, patient safety) the indicator relates to; name of the indicator; brief description of the indicator; numerator and denominator specifications; primary data source for the indicator. The Office of Management and Budget granted approval (0938-1153) for hospice data submission on January 3, 2012.
Vol.
-
No.
- -
2014
Hospices’ Use of EMRs for QAPI Programs
At the end of the voluntary reporting period on January 31, 2012, 911 hospices nationwide had submitted data about 6712 QIs used in their QAPI programs. These 911 hospices and the QIs they reported formed the analytic sample for the present study.
Analytic Approach We calculated the percentage of hospices that reported at least one EMR-based QI and the percentage of reported QIs created using each data source. QIs with EMRs as the data source are the focus of this study. Two authors independently classified the QIs for which the primary data source was EMRs into 10 quality domains (defined later), convened to check the level of agreement, discussed any disagreements, and came to a final disposition. We followed the quality domains that had been developed as part of the National Consensus Project and endorsed by the National Quality Forum (NQF).8,9 These quality domains are used by hospices and other stakeholders to drive the adoption of preferred practices and quality improvement. We slightly modified some domains and added a few domains to accommodate the QIs that we observed in the data. The added domains are consistent with the directions set by CMS in the Hospice Wage Index Final Rule for FY2012. The 10 patient care-related quality domains finally used in this study include the following: 1) physical symptom management, 2) patient safety, 3) structure and process of care, 4) patient and family preferences, 5) psychosocial management, 6) care coordination and transitions, 7) spiritual needs management, 8) grief bereavement and emotional support, 9) communication and education, and 10) patient and family experience/ratings of care and/or services. Each EMR-based QI was classified into one of the10 quality domains based on qualitative analyses of the name, brief description, and numerator and denominator statements of each QI. QIs that did not fit in any of these 10 quality domains were included in a separate domaindother.
Results Of the 911 reporting hospices, 878 submitted at least one QI during the voluntary
3
reporting period. Overall, 6712 QIs were reported. Table 1 summarizes the data sources for these QIs. EMR was identified as the most frequent data source, accounting for more than one-third (34.3%) of the reported QIs. About 13% of the QIs were based on paper medical records; about 30% were derived from a family survey or questionnaire. Less than 2% of the QIs were based on patient survey or questionnaire; almost 22% were based on other data sources such as incident report or log. Overall, about 70% of hospices (614 of 878, data not shown in tables) in our sample reported at least one EMR-based QI.
Quality Domains and Themes for EMR-Based QIs Table 2 summarizes the number of EMRbased QIs in each quality domain and the themes for the QIs in each domain. Physical symptom management was the largest quality domain; it included more than half (51.5%) of the EMR-based QIs in our data set. The vast majority of hospices with at least one EMR-based QI (89.3%) had QIs addressing physical symptom management. QIs in this quality domain addressed pain, dyspnea, bowel management, and nausea; the vast majority of indicators (773 or 65.3% of the physical symptom management QIs) addressed pain management. The most frequently reported pain management QIs addressed screening and/ or assessment of symptoms, initiation of interventions, reassessment of level of symptoms and patient comfort, and management of symptoms in the last week of life. Hospices infrequently reported EMR-based QIs that track incidence of uncontrolled symptoms (e.g., problem scores) or patient or family education about symptom management. Table 1 Data Sources for QIs Reported by Hospices During Voluntary Reporting Period QIs Reporteda Data Source EMR Paper medical record Family survey/questionnaire Patient survey/questionnaire Other Total
Number
Percent
2299 838 2027 105 1443 6712
34.3 12.5 30.2 1.6 21.5 100.0
QIs ¼ quality indicators; EMR ¼ electronic medical records. a Includes both EMR-based QIs and non-EMR-based QIs.
4
Zheng et al.
Vol.
-
No.
- -
2014
Table 2 Summary of EMR-Based QIs By Quality Domain
Quality Domain Physical symptom management
Number of Percent of Indicators Totala
Number of Hospices Reporting Percent of Totalb
1184
51.5
548
89.3
Patient safety
394
17.1
232
37.8
Structure and process of care
217
9.4
144
23.5
Patient and family preferences
151
6.6
124
20.2
Psychosocial management
119
5.2
91
14.8
Care coordination and transitions
49
2.1
45
7.3
Spiritual
46
2.0
45
7.3
Grief, bereavement, and emotional support
44
1.9
39
6.4
Themes Screening and/or assessment of symptoms - Effectiveness/outcomes of interventions on symptom control - Symptom management in last week of life - Ratings of comfort by patient - Tracking incidence of uncontrolled symptoms across care episodes - Patient/family education about symptom management - Falls tracking and/or prevention - Infections tracking and/or prevention - Medication safety Oxygen safetydrisk assessment and education - Compliance with Conditions of Participation - Visit frequency - Use of volunteer services - Advance directives/surrogate designation - Documenting patient/family preferences and goals of care - Meeting patient/family preferences and goals of care - Screening and/or assessment of symptoms - Effectiveness/outcomes of interventions on symptom control - Assessment and management of social support - Other psychological assessment or management - Care coordination among hospice care professionals - Care coordination between hospice and other professional care providers or settings - Transitions to other care settings - Transitions from other care settings - Screening/assessment/ management of spiritual needs/issues - Spiritual caredother (e.g., community clergy/spiritual care collaboration) - Grief and bereavement assessment, and care - Emotional care for patient/ family before and/or at time of death - Emotional care for family after the death - Culturally sensitive caregiving -
(Continued)
Vol.
-
No.
- -
2014
Hospices’ Use of EMRs for QAPI Programs
5
Table 2 Continued
Quality Domain
Number of Percent of Indicators Totala
Number of Hospices Reporting Percent of Totalb
0
Family education/ communication about the dying process - Patients/family participation in care planning - Use of a translator/interpreter - Communication and educationdother (e.g., durable medical equipment and one nurse in charge of patient care) - Lengths of stay - Distribution of terminal diagnosis - Live discharges - Quality of life NA
d
d
Communication and education
24
1.0
20
3.3
Other
71
3.1
48
7.8
Family experience/ratings of care and/or services Total
0
0
0
2299
100
614
Themes -
EMR ¼ electronic medical records; QIs ¼ quality indicators; NA ¼ not available. a The denominator for the percentages was the total number of EMR-based QIs. b The denominator for the percentages was the total number of hospices that reported at least one EMR-based QI.
We observed considerable variation in the specifications (i.e., measure denominator, numerator, and exclusions) of all quality domains, but it was most noteworthy in the domain of physical symptom management, particularly pain management. Several measures that address physical symptom management have been endorsed by the NQF10,11 and are commonly used by hospices; the variation in measure specifications we observed was, therefore, unexpected. For example, the NQF#0209 measures percent of patients who report being uncomfortable because of pain at initial assessment and who report pain brought to a comfortable level within 48 hours. Hospices reported at least six variants of this measure. These variants used patients’ comfort level at different time points to determine numerator, reversed the measure (i.e., measured the percentage of patients with an unacceptable level pain), or had different denominator exclusions (e.g., excluded patients unable to self-report pain information at the follow-up assessment). The second largest quality domain for the EMR-based QIs was patient safety; it included 394 (17.1%) of reported EMR-based QIs. About 37.8% of hospices with at least one EMR-based QI had QIs addressing patient safety. More than half of these QIs addressed
falls (tracking or prevention, or both): examples included falls with injury per 1000 patient-days and percentage of patients who had a plan of care specific to identified fall risks. The rest of the QIs in this domain addressed the following problems: 1) infections tracking and/or prevention (e.g., total number of infections per 1000 patient-days and percentage of patients screened for an infection), 2) medication safety (e.g., medication errors per 1000 patient-days, percentage of patient records in which medication reconciliation is completed at time of hospice admission and percentage of patients who have had teaching regarding medication administration and disposal), and 3) oxygen safety, risk assessment, and education (percentage of patients newly receiving oxygen with documentation of fire risk assessment and number of patients who had oxygen teaching done and documented when oxygen was initiated anytime after admission). Structure and process of care was the third largest quality domaind217 QIs (9.4% of reported EMR-based QIs). Almost one-quarter (23.5%) of hospices with at least one EMRbased QI had QIs addressing structure and process of care. QIs in this domain addressed hospices’ compliance with requirements of the Conditions of Participation (e.g.,
6
Zheng et al.
percentage of comprehensive assessments completed within five days of admission and percent of patients who have documentation of hospice aide supervision by the registered nurse every two weeks), visit frequency (e.g., average number of patient visits, clinical staff visit frequency, and visit frequency noted on the plan of care is adhered to or exceeded), and use of volunteer services (e.g., percentage of patients assigned a volunteer and percentage of volunteer hours). Patient and family preferences had 151 QIs (6.6% of reported QIs); about 20.2% of hospices with at least one EMR-based QI had QIs in this domain. QIs addressing advance directives or surrogate designation (e.g., documentation of advance directives or discussion, and percent of patients with contact information for surrogate decision maker) were the main elements of this domain. A few QIs addressed documenting patient or family preferences and goals of care with regard to other aspects of care, such as percentage of patients who have documentation of goals that are individualized and updated and percentage of care plans reviewed or revised every 14 days. A little less than half of QIs in this domain addressed meeting patient or family preferences and goals of care, for example, percentage of patients who avoided unwanted hospitalization, percentage of patients who received cardiopulmonary resuscitation and wished not to receive it, and percentage of patients who die in the setting of their choice. The psychosocial management domain had 119 QIs (5.2% of reported QIs). About 14.8% of hospices that reported at least one EMR-based QI had QIs in this domain. Most QIs in this domain addressed screening and/ or assessment and effectiveness or outcomes of interventions on symptom control for patients with anxiety and depression, mirroring the themes we found in the physical symptoms domain. Other QIs in this domain addressed assessment and management of social support. A small number of QIs addressed assessment or management of other psychological symptoms such as delirium and insomnia. The last four quality domainsdcare coordination and transitions, spiritual, grief bereavement and emotional support, and communication and educationdeach had few reported EMRbased QIs. The themes in these domains are
Vol.
-
No.
- -
2014
presented in Table 2. A total of 71 EMR-based QIs addressed other aspects of care such as lengths of stay, distribution of terminal diagnosis (e.g., percent of cancer patients), live discharges, and quality of life. No hospices reported EMRbased QIs addressing the quality domain of family experience or ratings of care or services.
Discussion This is the first study that examines the frequency and content of EMR-based QIs used by hospices specifically for quality assessment and improvement purposes. Given that the Affordable Care Act requires the development of a national quality reporting system for hospices, examining EMR use by hospices for creating QIs is timely and important. Our analyses provide information about the ability of hospices to gather and submit quality data and extend the current literature on the use of electronic data sources in the context of quality improvement for hospices. Our findings regarding the patient care quality domains that hospices monitor with EMR-based QIs as part of their QAPI programs can help guide policymakers’ decisions about the future development of hospice quality reporting requirements. The findings also provide insight for EMR vendors who want to improve their EMR systems to better facilitate hospices’ quality improvement efforts. EMRs were the most frequent data source for the QIs reported during the voluntary reporting period. Most hospices that participated in the voluntary data reporting used EMRs as a data source for at least one QI in their QAPI programs. Our results show that hospices were using EMR-based QIs to measure aspects of quality across all domains of care, except for family experience/ratings of care and/or services. QIs in this domain were typically based on patient or family surveys, particularly the Family Evaluation of Hospice Care (FEHC) survey (a postdeath survey designed by the National Hospice and Palliative Care Organization to yield actionable information that reflects the quality of hospice care delivery from the perspective of family caregivers).12 The FEHC survey is a standardized quality data collection tool and can provide important information
Vol.
-
No.
- -
2014
Hospices’ Use of EMRs for QAPI Programs
for national hospice quality evaluation in addition to EMR-based QIs. Many of the reported EMR-based QIs were homegrown by individual hospices and addressed specific quality concerns for that individual hospice; they are not used for comparing quality results with other hospices. By contrast, many other QIs reported by hospices were well-known measures developed under previous CMS-funded projects.13,14 These QIs target the same quality themes across providers and, therefore, allow potential quality comparison and benchmarking. In February 2012, the NQF endorsed 14 quality measures of palliative and end-of-life care, addressing a wide range of care areas and concerns, including pain management, psychosocial needs, care transitions, and experience of care.10 Endorsement of these quality measures, particularly the nonpain quality measures (e.g., patients treated with an opioid who are given a bowel regimen, dyspnea screening, and treatment preferences), may increasingly result in the use of more standardized measures in hospices’ QAPI programs. However, even among hospices that reported that they already use these standardized measures, we found a large amount of variation in the QI specifications reported by hospice providers. The variations in the QI specifications that hospices reported (e.g., measure denominator, numerator, exclusions, and time window) pose challenges to quality comparison across hospices. In the future, implementing a standardized patient-level data collection tool and the associated training and support for hospices may establish the consistency needed for nationwide hospice quality comparisons. This study had several limitations. First, the study included hospices that voluntarily reported information about their QAPI programs during the voluntary reporting period. According to CMS’s Quality Improvement Evaluation System database, 3790 Medicarecertified hospices were required to submit data for fiscal year 2014. Thus, our study hospices represented roughly a quarter of hospices nationwide assuming a similar number of hospices nationwide during our study period (Quarter 4, 2011). Hospices that chose to report voluntarily may have betterdeveloped QAPI programs; they also may be more likely to have used EMRs as a data source
7
for their QIs. Therefore, our findings may overestimate the use of EMRs as a data source for QIs. Future analysis, linking hospice-reported quality data with hospice characteristics (e.g., provider type, for-profit/ not-for-profit status, and facility daily census), may provide useful information about the generalizability of our findings. Second, the data entry system allowed hospices to select only one primary data source for each QI reported. Some EMR-based QIs in our data set required supplemental data sources. For example, one hospice reported that they use dyspnea symptom management as a QI. Data for this QI were obtained from EMRs and responses to the FEHC survey. EMRs provided information on comprehensive respiratory assessments, patient education on dyspnea, and interventions; the FEHC provided supplemental information on patient education on breathing problems and treatments for breathing. Finally, among hospices that did not report using EMR-based QIs in their QAPI programs, our data did not allow us to differentiate between hospices that did not have an EMR system in place and hospices that had an EMR system but did not use the data to create the QIs they used in their QAPI programs. Future studies may be needed to differentiate between the two causes. In conclusion, this study reveals that hospices reported a wide variety of EMR-based QIs in their QAPI programs, addressing nine of 10 quality domains. These findings present promising opportunities for future EMRbased quality reporting with regard to various quality domains. If a standardized patientlevel data collection tool is implemented and measures are specified based on the standardized data, more standardization may be seen among EMR vendors’ data input systems and the reports that these systems produce. EMRbased QIs could form the basis of nationwide quality reporting and comparison if there is standard EMR-generated quality data reporting across hospices.
Disclosures and Acknowledgments The authors of this study are employees of RTI International, an independent and nonprofit research institute. The authors’
8
Zheng et al.
research and writing activities were funded by an RTI internal professional development award. The data used in the study were collected under an RTI contract to the CMS to develop a quality reporting system for inpatient rehabilitation facilities, long-term care hospitals, and hospice providers (contract no.: HHSM-500-2008-00021I). The CMS reviewed and approved this study before submission.
References 1. Medicare Payment Advisory Commission. Hospice services. Report to the Congress: Medicare Payment Policy 2012. Washington, DC: Medicare Payment Advisory Commission, 2012:281e308. 2. U.S. Government. Patient Protection and Affordable Care Act (ACA), section 3004. Washington, DC: U.S. Government, 2010. 3. Centers for Medicare & Medicaid Services. Medicare program: hospice wage index for fiscal year 2012. Fed Regist 2011;76:47302e47352. 4. Resnick HE, Alwan M. Use of health information technology in home health and hospice agencies: United States, 2007. J Am Med Inform Assoc 2010;17:389e395. 5. Caffrey C, Sengupta A, Moss A, HarrisKojetlin L, Valverde R. Home health care and discharged hospice care patients: United States, 2000 and 2007. Hyattsville, MD: National Center for Health Statistics, 2011. 6. Bercovitz A, Sengupta M, Jamison P. Electronic medical record adoption and use in home health and hospice. NCHS Data Brief no. 45. Hyattsville, MD: National Center for Health Statistics, 2010.
Vol.
-
No.
- -
2014
7. Cagle JG, Rokoske FS, Durham D, et al. Use of electronic documentation for quality improvement in hospice. Am J Med Qual 2012;27:282e290. 8. National Consensus Project for Quality Palliative Care. Clinical practice guidelines for quality palliative care, 2nd ed. 2009. Available from http://www.nationalconsensusproject.org/guideline. pdf. Accessed December 14, 2012. 9. National Quality Forum. Palliative care and endof-life caredA consensus report. 2012. Available from http://www.qualityforum.org/Projects/Palliative_ Care_and_End-of-Life_Care.aspx. Accessed December 12, 2012. 10. National Quality Forum. NQF endorses palliative and end-of-life care measures. 2012. Available from http://www.qualityforum.org/News_And_ Resources/Press_Releases/2012/NQF_Endorses_ Palliative_and_End-of-Life_Care_Measures.aspx. Accessed November 7, 2012. 11. National Quality Forum. NQF#0209 Comfortable dying: pain brought to a comfortable level within 48 hours of initial assessment. 2012. Available from www.qualityforum.org. Accessed November 7, 2012. 12. National Hospice and Palliative Care Organization. Family Evaluation of Hospice Care (FEHC) Survey. 2012. Available from http://www.nhpco. org/i4a/pages/index.cfm?pageid¼6135. Accessed November 7, 2012. 13. Carolinas Center for Medical Excellence. Hospice PEACE project. 2008. Available from http:// www.thecarolinascenter.org/default.aspx?pageid¼ 24. Accessed November 7, 2012. 14. IPRO Hospice AIM Project Team. Hospice Assessment Intervention and Measurement (AIM) Toolkit. 2011. Available from http://www.ipro.org/ index/hospice-aim. Accessed November 6, 2012.
-
No.
- -
2014
Journal of Pain and Symptom Management 1
Original Article
Hospices’ Use of Electronic Medical Records for Quality Assessment and Performance Improvement Programs Nan Tracy Zheng, PhD, Franziska S. Rokoske, PT, MS, M. Alexis Kirk, BA, Brieanne Lyda-McDonald, MSPH, and Shulamit L. Bernard, PhD Division of Health Services and Social Policy Research (N.T.Z.), RTI International, Waltham, Massachusetts; and Division of Health Services and Social Policy Research (F.S.R., M.A.K., B.L.-M., S.L.B.), RTI International, Research Triangle Park, North Carolina, USA
Abstract Context. Electronic medical records (EMRs) are increasingly viewed as essential tools for quality assurance and improvement in many care settings, but little is known about the use of EMRs by hospices in their quality assessment and performance improvement (QAPI) programs. Objectives. To examine the data sources hospices use to create quality indicators (QIs) used in their QAPI programs and to examine the domains of EMR-based QIs. Methods. We used self-reported QIs (description, numerator, and denominator) from 911 hospices nationwide that participated in the Centers for Medicare & Medicaid Services nationwide hospice voluntary reporting period. The data reflected QIs that hospices used for their internal QAPI programs between October 1 and December 31, 2011. We used the primary data sources for QIs reported by hospices and analyzed EMR-based QIs in terms of the quality domains and themes addressed. Results. EMRs were the most frequent data source for the QIs reported, followed by family survey and paper medical record. Physical symptom management was the largest quality domaindincluded in 51.5% of the reported EMR-based QIsdfollowed by patient safety and structure and process of care. Conclusion. Most participating hospices use EMRs for retrieving items needed for QI calculations. EMR-based QIs address various quality domains and themes. Our findings present opportunities for potential future reporting of EMR-based quality data. J Pain Symptom Manage 2014;-:-e-. Ó 2014 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Electronic medical record, electronic health record, hospice quality indicators, quality reporting
Address correspondence to: Nan Tracy Zheng, PhD, Aging, Disability and Long Term Care, Division of Health Services and Social Policy Research, RTI Ó 2014 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved.
International, 1440 Main Street, Suite 310, Waltham, MA 02451-1623, USA. E-mail: ntzheng@ rti.org 0885-3924/$ - see front matter http://dx.doi.org/10.1016/j.jpainsymman.2013.11.010
2
Zheng et al.
Introduction Medicare hospice services include palliation, comfort, and support services for Medicare beneficiaries with a life expectancy of six months or less and their families. In 2009, about 1.1 million beneficiaries (more than 40% of Medicare decedents in that year) received hospice services.1 Medicare hospice spending more than tripled between 2000 ($2.9 billion) and 2007 ($10 billion) and increased to $12 billion in 2009. At the same time, the number of hospice providers that are certified by the Centers for Medicare & Medicaid Services (CMS) has increased rapidly, particularly for-profit hospices, which increased by 150% between 2000 and 2010.1 Despite the high hospice utilization and the proliferation of hospice providers, the public and policymakers know relatively little about the quality of care provided by hospices. Developing, implementing, and maintaining an effective, ongoing, hospice-wide, data-driven quality assessment and performance improvement (QAPI) program has been a requirement in the Conditions of Participation since 2008, but whether hospices implement QAPI programs, and how many and what type of quality indicators (QIs) are included in the QAPI programs, are largely unknown. The Affordable Care Act required the U.S. Department of Health and Human Services to establish a Hospice Quality Reporting Program.2 CMS initiated a voluntary reporting period beginning in January 2012 for hospices nationwide to report the QIs that they used in their QAPI programs between October 1 and December 31, 2011.3 The purpose of this voluntary reporting period was to help CMS design the data collection approach for mandatory quality reporting starting in 2013. The data reported by hospices during the voluntary reporting period also provided the first opportunity to identify the QIs that they are using to meet QAPI requirements and the data sources for constructing their QIs. Electronic medical records (EMRs) are increasingly viewed as essential tools for quality assurance and improvement in a variety of care settings.4 The 2007 National Home Health and Hospice Care Survey found that approximately 49% of hospice agencies and 63% of agencies that provide both home health and hospice
Vol.
-
No.
- -
2014
care were using EMRs.5 The most common EMR functionalities or components that providers use are for recording patient demographics and clinical notes.6 Only one study investigated the use of EMRs specifically for the purposes of quality measurement or improvement in hospices. Cagle et al.7 reported on the use of electronic data (including EMRs) for QAPI purposes. Their survey of more than 650 hospice providers showed that larger, home-health agency-based, nonprofit, mixed urban/rural hospices had the highest rates of use of electronic data for quality measurement or improvement. Although this survey provided evidence that hospices were using electronic data as part of their quality measurement efforts, it did not collect data on what QIs hospices created using electronic data. The purposes of this study were to examine the data sources that hospices use to create individual QIs and to examine the domains of patient care addressed by EMR-based QIs.
Methods Data and Sample In late 2011, all hospices nationwide were made aware of their eligibility to participate in the voluntary reporting period through the CMS Hospice Wage Index Final Rule for FY2012 (CMS-1355-F, August 2011), CMS Open Door Forums, information posted on the CMS Hospice Quality Reporting Program Web site, and provider association newsblasts. A Web-based data collection form was created for hospices to submit data during the voluntary reporting period. Hospices were asked to enter details about the QIs they used if they had at least one patient care-related QI in their QAPI program in Quarter 4 of 2011 (October 1eDecember 31). For each QI, hospices provided the following information by either selecting from a drop-down menu or entering free text: patient care topic (e.g., pain management, psychosocial assessment, patient safety) the indicator relates to; name of the indicator; brief description of the indicator; numerator and denominator specifications; primary data source for the indicator. The Office of Management and Budget granted approval (0938-1153) for hospice data submission on January 3, 2012.
Vol.
-
No.
- -
2014
Hospices’ Use of EMRs for QAPI Programs
At the end of the voluntary reporting period on January 31, 2012, 911 hospices nationwide had submitted data about 6712 QIs used in their QAPI programs. These 911 hospices and the QIs they reported formed the analytic sample for the present study.
Analytic Approach We calculated the percentage of hospices that reported at least one EMR-based QI and the percentage of reported QIs created using each data source. QIs with EMRs as the data source are the focus of this study. Two authors independently classified the QIs for which the primary data source was EMRs into 10 quality domains (defined later), convened to check the level of agreement, discussed any disagreements, and came to a final disposition. We followed the quality domains that had been developed as part of the National Consensus Project and endorsed by the National Quality Forum (NQF).8,9 These quality domains are used by hospices and other stakeholders to drive the adoption of preferred practices and quality improvement. We slightly modified some domains and added a few domains to accommodate the QIs that we observed in the data. The added domains are consistent with the directions set by CMS in the Hospice Wage Index Final Rule for FY2012. The 10 patient care-related quality domains finally used in this study include the following: 1) physical symptom management, 2) patient safety, 3) structure and process of care, 4) patient and family preferences, 5) psychosocial management, 6) care coordination and transitions, 7) spiritual needs management, 8) grief bereavement and emotional support, 9) communication and education, and 10) patient and family experience/ratings of care and/or services. Each EMR-based QI was classified into one of the10 quality domains based on qualitative analyses of the name, brief description, and numerator and denominator statements of each QI. QIs that did not fit in any of these 10 quality domains were included in a separate domaindother.
Results Of the 911 reporting hospices, 878 submitted at least one QI during the voluntary
3
reporting period. Overall, 6712 QIs were reported. Table 1 summarizes the data sources for these QIs. EMR was identified as the most frequent data source, accounting for more than one-third (34.3%) of the reported QIs. About 13% of the QIs were based on paper medical records; about 30% were derived from a family survey or questionnaire. Less than 2% of the QIs were based on patient survey or questionnaire; almost 22% were based on other data sources such as incident report or log. Overall, about 70% of hospices (614 of 878, data not shown in tables) in our sample reported at least one EMR-based QI.
Quality Domains and Themes for EMR-Based QIs Table 2 summarizes the number of EMRbased QIs in each quality domain and the themes for the QIs in each domain. Physical symptom management was the largest quality domain; it included more than half (51.5%) of the EMR-based QIs in our data set. The vast majority of hospices with at least one EMR-based QI (89.3%) had QIs addressing physical symptom management. QIs in this quality domain addressed pain, dyspnea, bowel management, and nausea; the vast majority of indicators (773 or 65.3% of the physical symptom management QIs) addressed pain management. The most frequently reported pain management QIs addressed screening and/ or assessment of symptoms, initiation of interventions, reassessment of level of symptoms and patient comfort, and management of symptoms in the last week of life. Hospices infrequently reported EMR-based QIs that track incidence of uncontrolled symptoms (e.g., problem scores) or patient or family education about symptom management. Table 1 Data Sources for QIs Reported by Hospices During Voluntary Reporting Period QIs Reporteda Data Source EMR Paper medical record Family survey/questionnaire Patient survey/questionnaire Other Total
Number
Percent
2299 838 2027 105 1443 6712
34.3 12.5 30.2 1.6 21.5 100.0
QIs ¼ quality indicators; EMR ¼ electronic medical records. a Includes both EMR-based QIs and non-EMR-based QIs.
4
Zheng et al.
Vol.
-
No.
- -
2014
Table 2 Summary of EMR-Based QIs By Quality Domain
Quality Domain Physical symptom management
Number of Percent of Indicators Totala
Number of Hospices Reporting Percent of Totalb
1184
51.5
548
89.3
Patient safety
394
17.1
232
37.8
Structure and process of care
217
9.4
144
23.5
Patient and family preferences
151
6.6
124
20.2
Psychosocial management
119
5.2
91
14.8
Care coordination and transitions
49
2.1
45
7.3
Spiritual
46
2.0
45
7.3
Grief, bereavement, and emotional support
44
1.9
39
6.4
Themes Screening and/or assessment of symptoms - Effectiveness/outcomes of interventions on symptom control - Symptom management in last week of life - Ratings of comfort by patient - Tracking incidence of uncontrolled symptoms across care episodes - Patient/family education about symptom management - Falls tracking and/or prevention - Infections tracking and/or prevention - Medication safety Oxygen safetydrisk assessment and education - Compliance with Conditions of Participation - Visit frequency - Use of volunteer services - Advance directives/surrogate designation - Documenting patient/family preferences and goals of care - Meeting patient/family preferences and goals of care - Screening and/or assessment of symptoms - Effectiveness/outcomes of interventions on symptom control - Assessment and management of social support - Other psychological assessment or management - Care coordination among hospice care professionals - Care coordination between hospice and other professional care providers or settings - Transitions to other care settings - Transitions from other care settings - Screening/assessment/ management of spiritual needs/issues - Spiritual caredother (e.g., community clergy/spiritual care collaboration) - Grief and bereavement assessment, and care - Emotional care for patient/ family before and/or at time of death - Emotional care for family after the death - Culturally sensitive caregiving -
(Continued)
Vol.
-
No.
- -
2014
Hospices’ Use of EMRs for QAPI Programs
5
Table 2 Continued
Quality Domain
Number of Percent of Indicators Totala
Number of Hospices Reporting Percent of Totalb
0
Family education/ communication about the dying process - Patients/family participation in care planning - Use of a translator/interpreter - Communication and educationdother (e.g., durable medical equipment and one nurse in charge of patient care) - Lengths of stay - Distribution of terminal diagnosis - Live discharges - Quality of life NA
d
d
Communication and education
24
1.0
20
3.3
Other
71
3.1
48
7.8
Family experience/ratings of care and/or services Total
0
0
0
2299
100
614
Themes -
EMR ¼ electronic medical records; QIs ¼ quality indicators; NA ¼ not available. a The denominator for the percentages was the total number of EMR-based QIs. b The denominator for the percentages was the total number of hospices that reported at least one EMR-based QI.
We observed considerable variation in the specifications (i.e., measure denominator, numerator, and exclusions) of all quality domains, but it was most noteworthy in the domain of physical symptom management, particularly pain management. Several measures that address physical symptom management have been endorsed by the NQF10,11 and are commonly used by hospices; the variation in measure specifications we observed was, therefore, unexpected. For example, the NQF#0209 measures percent of patients who report being uncomfortable because of pain at initial assessment and who report pain brought to a comfortable level within 48 hours. Hospices reported at least six variants of this measure. These variants used patients’ comfort level at different time points to determine numerator, reversed the measure (i.e., measured the percentage of patients with an unacceptable level pain), or had different denominator exclusions (e.g., excluded patients unable to self-report pain information at the follow-up assessment). The second largest quality domain for the EMR-based QIs was patient safety; it included 394 (17.1%) of reported EMR-based QIs. About 37.8% of hospices with at least one EMR-based QI had QIs addressing patient safety. More than half of these QIs addressed
falls (tracking or prevention, or both): examples included falls with injury per 1000 patient-days and percentage of patients who had a plan of care specific to identified fall risks. The rest of the QIs in this domain addressed the following problems: 1) infections tracking and/or prevention (e.g., total number of infections per 1000 patient-days and percentage of patients screened for an infection), 2) medication safety (e.g., medication errors per 1000 patient-days, percentage of patient records in which medication reconciliation is completed at time of hospice admission and percentage of patients who have had teaching regarding medication administration and disposal), and 3) oxygen safety, risk assessment, and education (percentage of patients newly receiving oxygen with documentation of fire risk assessment and number of patients who had oxygen teaching done and documented when oxygen was initiated anytime after admission). Structure and process of care was the third largest quality domaind217 QIs (9.4% of reported EMR-based QIs). Almost one-quarter (23.5%) of hospices with at least one EMRbased QI had QIs addressing structure and process of care. QIs in this domain addressed hospices’ compliance with requirements of the Conditions of Participation (e.g.,
6
Zheng et al.
percentage of comprehensive assessments completed within five days of admission and percent of patients who have documentation of hospice aide supervision by the registered nurse every two weeks), visit frequency (e.g., average number of patient visits, clinical staff visit frequency, and visit frequency noted on the plan of care is adhered to or exceeded), and use of volunteer services (e.g., percentage of patients assigned a volunteer and percentage of volunteer hours). Patient and family preferences had 151 QIs (6.6% of reported QIs); about 20.2% of hospices with at least one EMR-based QI had QIs in this domain. QIs addressing advance directives or surrogate designation (e.g., documentation of advance directives or discussion, and percent of patients with contact information for surrogate decision maker) were the main elements of this domain. A few QIs addressed documenting patient or family preferences and goals of care with regard to other aspects of care, such as percentage of patients who have documentation of goals that are individualized and updated and percentage of care plans reviewed or revised every 14 days. A little less than half of QIs in this domain addressed meeting patient or family preferences and goals of care, for example, percentage of patients who avoided unwanted hospitalization, percentage of patients who received cardiopulmonary resuscitation and wished not to receive it, and percentage of patients who die in the setting of their choice. The psychosocial management domain had 119 QIs (5.2% of reported QIs). About 14.8% of hospices that reported at least one EMR-based QI had QIs in this domain. Most QIs in this domain addressed screening and/ or assessment and effectiveness or outcomes of interventions on symptom control for patients with anxiety and depression, mirroring the themes we found in the physical symptoms domain. Other QIs in this domain addressed assessment and management of social support. A small number of QIs addressed assessment or management of other psychological symptoms such as delirium and insomnia. The last four quality domainsdcare coordination and transitions, spiritual, grief bereavement and emotional support, and communication and educationdeach had few reported EMRbased QIs. The themes in these domains are
Vol.
-
No.
- -
2014
presented in Table 2. A total of 71 EMR-based QIs addressed other aspects of care such as lengths of stay, distribution of terminal diagnosis (e.g., percent of cancer patients), live discharges, and quality of life. No hospices reported EMRbased QIs addressing the quality domain of family experience or ratings of care or services.
Discussion This is the first study that examines the frequency and content of EMR-based QIs used by hospices specifically for quality assessment and improvement purposes. Given that the Affordable Care Act requires the development of a national quality reporting system for hospices, examining EMR use by hospices for creating QIs is timely and important. Our analyses provide information about the ability of hospices to gather and submit quality data and extend the current literature on the use of electronic data sources in the context of quality improvement for hospices. Our findings regarding the patient care quality domains that hospices monitor with EMR-based QIs as part of their QAPI programs can help guide policymakers’ decisions about the future development of hospice quality reporting requirements. The findings also provide insight for EMR vendors who want to improve their EMR systems to better facilitate hospices’ quality improvement efforts. EMRs were the most frequent data source for the QIs reported during the voluntary reporting period. Most hospices that participated in the voluntary data reporting used EMRs as a data source for at least one QI in their QAPI programs. Our results show that hospices were using EMR-based QIs to measure aspects of quality across all domains of care, except for family experience/ratings of care and/or services. QIs in this domain were typically based on patient or family surveys, particularly the Family Evaluation of Hospice Care (FEHC) survey (a postdeath survey designed by the National Hospice and Palliative Care Organization to yield actionable information that reflects the quality of hospice care delivery from the perspective of family caregivers).12 The FEHC survey is a standardized quality data collection tool and can provide important information
Vol.
-
No.
- -
2014
Hospices’ Use of EMRs for QAPI Programs
for national hospice quality evaluation in addition to EMR-based QIs. Many of the reported EMR-based QIs were homegrown by individual hospices and addressed specific quality concerns for that individual hospice; they are not used for comparing quality results with other hospices. By contrast, many other QIs reported by hospices were well-known measures developed under previous CMS-funded projects.13,14 These QIs target the same quality themes across providers and, therefore, allow potential quality comparison and benchmarking. In February 2012, the NQF endorsed 14 quality measures of palliative and end-of-life care, addressing a wide range of care areas and concerns, including pain management, psychosocial needs, care transitions, and experience of care.10 Endorsement of these quality measures, particularly the nonpain quality measures (e.g., patients treated with an opioid who are given a bowel regimen, dyspnea screening, and treatment preferences), may increasingly result in the use of more standardized measures in hospices’ QAPI programs. However, even among hospices that reported that they already use these standardized measures, we found a large amount of variation in the QI specifications reported by hospice providers. The variations in the QI specifications that hospices reported (e.g., measure denominator, numerator, exclusions, and time window) pose challenges to quality comparison across hospices. In the future, implementing a standardized patient-level data collection tool and the associated training and support for hospices may establish the consistency needed for nationwide hospice quality comparisons. This study had several limitations. First, the study included hospices that voluntarily reported information about their QAPI programs during the voluntary reporting period. According to CMS’s Quality Improvement Evaluation System database, 3790 Medicarecertified hospices were required to submit data for fiscal year 2014. Thus, our study hospices represented roughly a quarter of hospices nationwide assuming a similar number of hospices nationwide during our study period (Quarter 4, 2011). Hospices that chose to report voluntarily may have betterdeveloped QAPI programs; they also may be more likely to have used EMRs as a data source
7
for their QIs. Therefore, our findings may overestimate the use of EMRs as a data source for QIs. Future analysis, linking hospice-reported quality data with hospice characteristics (e.g., provider type, for-profit/ not-for-profit status, and facility daily census), may provide useful information about the generalizability of our findings. Second, the data entry system allowed hospices to select only one primary data source for each QI reported. Some EMR-based QIs in our data set required supplemental data sources. For example, one hospice reported that they use dyspnea symptom management as a QI. Data for this QI were obtained from EMRs and responses to the FEHC survey. EMRs provided information on comprehensive respiratory assessments, patient education on dyspnea, and interventions; the FEHC provided supplemental information on patient education on breathing problems and treatments for breathing. Finally, among hospices that did not report using EMR-based QIs in their QAPI programs, our data did not allow us to differentiate between hospices that did not have an EMR system in place and hospices that had an EMR system but did not use the data to create the QIs they used in their QAPI programs. Future studies may be needed to differentiate between the two causes. In conclusion, this study reveals that hospices reported a wide variety of EMR-based QIs in their QAPI programs, addressing nine of 10 quality domains. These findings present promising opportunities for future EMRbased quality reporting with regard to various quality domains. If a standardized patientlevel data collection tool is implemented and measures are specified based on the standardized data, more standardization may be seen among EMR vendors’ data input systems and the reports that these systems produce. EMRbased QIs could form the basis of nationwide quality reporting and comparison if there is standard EMR-generated quality data reporting across hospices.
Disclosures and Acknowledgments The authors of this study are employees of RTI International, an independent and nonprofit research institute. The authors’
8
Zheng et al.
research and writing activities were funded by an RTI internal professional development award. The data used in the study were collected under an RTI contract to the CMS to develop a quality reporting system for inpatient rehabilitation facilities, long-term care hospitals, and hospice providers (contract no.: HHSM-500-2008-00021I). The CMS reviewed and approved this study before submission.
References 1. Medicare Payment Advisory Commission. Hospice services. Report to the Congress: Medicare Payment Policy 2012. Washington, DC: Medicare Payment Advisory Commission, 2012:281e308. 2. U.S. Government. Patient Protection and Affordable Care Act (ACA), section 3004. Washington, DC: U.S. Government, 2010. 3. Centers for Medicare & Medicaid Services. Medicare program: hospice wage index for fiscal year 2012. Fed Regist 2011;76:47302e47352. 4. Resnick HE, Alwan M. Use of health information technology in home health and hospice agencies: United States, 2007. J Am Med Inform Assoc 2010;17:389e395. 5. Caffrey C, Sengupta A, Moss A, HarrisKojetlin L, Valverde R. Home health care and discharged hospice care patients: United States, 2000 and 2007. Hyattsville, MD: National Center for Health Statistics, 2011. 6. Bercovitz A, Sengupta M, Jamison P. Electronic medical record adoption and use in home health and hospice. NCHS Data Brief no. 45. Hyattsville, MD: National Center for Health Statistics, 2010.
Vol.
-
No.
- -
2014
7. Cagle JG, Rokoske FS, Durham D, et al. Use of electronic documentation for quality improvement in hospice. Am J Med Qual 2012;27:282e290. 8. National Consensus Project for Quality Palliative Care. Clinical practice guidelines for quality palliative care, 2nd ed. 2009. Available from http://www.nationalconsensusproject.org/guideline. pdf. Accessed December 14, 2012. 9. National Quality Forum. Palliative care and endof-life caredA consensus report. 2012. Available from http://www.qualityforum.org/Projects/Palliative_ Care_and_End-of-Life_Care.aspx. Accessed December 12, 2012. 10. National Quality Forum. NQF endorses palliative and end-of-life care measures. 2012. Available from http://www.qualityforum.org/News_And_ Resources/Press_Releases/2012/NQF_Endorses_ Palliative_and_End-of-Life_Care_Measures.aspx. Accessed November 7, 2012. 11. National Quality Forum. NQF#0209 Comfortable dying: pain brought to a comfortable level within 48 hours of initial assessment. 2012. Available from www.qualityforum.org. Accessed November 7, 2012. 12. National Hospice and Palliative Care Organization. Family Evaluation of Hospice Care (FEHC) Survey. 2012. Available from http://www.nhpco. org/i4a/pages/index.cfm?pageid¼6135. Accessed November 7, 2012. 13. Carolinas Center for Medical Excellence. Hospice PEACE project. 2008. Available from http:// www.thecarolinascenter.org/default.aspx?pageid¼ 24. Accessed November 7, 2012. 14. IPRO Hospice AIM Project Team. Hospice Assessment Intervention and Measurement (AIM) Toolkit. 2011. Available from http://www.ipro.org/ index/hospice-aim. Accessed November 6, 2012.
