The whys and hows of going beyond CMS measures in outcome analytics
The emerging model of value-based care delivery stated loud and clear that quality is not in quantity. CMS plans are that 30 percent of fee-for-service Medicare payments will be changed to value-based payments via Accountable Care Organizations (ACOs) and bundling by the end of 2016. Moreover, 55 percent of FFS payments are anticipated to transform into quality-based payments through pay-for-performance (P4P) programs by the end of 2016 either.
This means that in order to provide valid information on reports to CMS for appropriate performance evaluation and forthcoming incentives, caregivers will need to gather yet more data on outcomes and analyze it.
More reasons to expand and refine outcomes in data analytics
Beyond this, the outcomes-focused healthcare data analytics allows providers to achieve a number of important performance goals, such as:
- Finding and fixing problems in internal clinical processes
- Predicting and eliminating patient satisfaction gaps
- Making comparisons between facilities, doctors and nurses to spot ‘pacesetters’ (to extract best practices) and ‘underperformers’ (to fix the situation)
However, caregivers can’t reach these goals by using only the CMS measures, since they are fragmented and too general, trying to cover various diseases with the same measures.
Why CMS measures aren’t enough for health outcome analytics
The CMS quality measures have been created on a consensual basis. Committees, agencies and other organizations such as NCQA, AHRQ, TJC and American College of Cardiology have submitted their suggestions. Then the Core Quality Measure Collaborative collectively decided on whether the submitted measures were applicable. The resulting sets of measures are diverse and disconnected, probably due to the absence of a core framework with a clear structure.
Therefore, caregivers might misuse the CMS measures for internal performance optimization, as the outcomes-focused criteria bring in an incomplete picture and selective measures.
According to CMS, the following 7 clinical areas contain the core measure sets:
- Accountable Care Organizations (ACOs), Patient Centered Medical Homes (PCMH), and Primary Care
- HIV and Hepatitis C
- Medical Oncology
- Obstetrics and Gynecology
Sadly, these sets mostly concern clinical processes, so outcomes-focused measures are scanty. Still, there is hope about more outcomes-bound criteria to come, as each document on a particular core measure set includes a list of future areas to be considered:
For ACOs, PCMH and Primary Care:
- Health-related quality of life
- Patient-reported outcomes (PROs)
- Pain management measures
- PROs for asthma exacerbations
- Outpatient – symptom control or change in symptoms
- Renal function measures for hypertension
- Rehabilitation measures
- Mental health measures following cardiovascular events
- Symptom management measures
- GERD and cirrhosis measures
- Patient safety measure: 7-day risk-standardized hospital visit rate after outpatient colonoscopy
- Pain control
- Functional status or quality of life
- Disease-free survival for X number of years
- Patient experience / PROs for the level of pain experienced by a patient
- Length of stay
- Return to surgery (infection, frozen joint, etc.)
- Adverse events surrounding surgery (post-operative cellulitis, pneumonia, etc.)
- Functional status measures for patients undergoing orthopedic surgery
The HIV and Hepatitis C area along with Obstetrics and Gynecology don’t have any outcomes-focused suggestions for the future and they don’t currently focus on outcomes-related measures.
Conclusion: Dividing quality measures by departments seems to be the right approach, yet CMS may face certain obstacles in achieving their ambitious goals by the end of 2016 due to a rather insufficient set of outcome measures. On their side, caregivers will have to adopt the CMS measures for reporting anyway, yet those measures are still insufficient to meet the internal clinical objectives we defined above.
Beyond CMS: AHRQ, NIH and Mayo Clinic review
AHRQ’s National Quality Measures Clearinghouse (NQMC) offers a wide range of quality measures for healthcare data analytics for providers. Caregivers can filter them up using tabs or by typing relevant keywords; there is also a possibility to compare them. Currently, there are 246 different outcomes-focused measures supported by rationale, data collection guidelines, measure computation and more valuable information.
With this volume of measures available, there is only a limited number of diseases covered, thus these sets provide caregivers only with a rather fragmented picture of their care quality.
For example, diabetes is more or less covered with a set of 57 measures, including such valuable criteria as:
- Percentage of patients with a blood pressure reading less than 130 / 80 in the last 12 months.
- Percentage of patients from 18 to 75 years with type 1 or type 2 diabetes whose condition was optimally managed during the measurement period.
- Percentage of patients from 18 to 75 years with type 1 or type 2 diabetes whose most recent hemoglobin A1c (HbA1c) level is greater than 9.0% (poorly controlled).
Now to non-chronic conditions. Pneumonia, for instance, isn’t sufficiently covered. With only 3 measures related to it, two are on mortality:
- Hospital 30-day, all-cause, risk-standardized mortality rate (RSMR) following pneumonia hospitalization.
- Ratio of episodes of ventilator-associated pneumonia to days of invasive mechanical ventilation (MV).
Conclusion: Even when NQMC offers more disease-specific outcome measures for data analytics, there is a gap between highly covered and disregarded conditions.
NIH’s HealthMeasures is a web resource, combining 4 measurement systems – PROMIS®, NIH Toolbox® (these two target general population with health conditions), Neuro-QoL (targets neurological patients) and ASCQ-MeSM (targets individuals with sickle cell disease).
Overall, these systems include a wide range of self-reported and proxy-reported measures along with performance tests of cognitive, motor, and sensory functions to illustrate patients’ health outcomes for the functions, symptoms and sensations. Caregivers can access these outcome measures from all the systems simultaneously, using the ‘Search & View Measures’ feature.
There are 6 types of filters available for caregivers:
- Age (adult, pediatric, proxy report for pediatric)
- Category (mental, physical and social health along with global plus multiple health)
- Domain (alcohol use, anger, anxiety / fear, communication, depression / sadness, end of life concerns, stress, smoking, pain, fatigue, dyspnea, motor function, gastrointestinal and more)
- Measure type (fixed length short form, computer adaptive test / item bank, battery / profile, performance test)
- Measurement system (PROMIS®, Neuro-QoL, ASCQ-MeSM and NIH Toolbox®)
HealthMeasures also offers a few guidelines on applying the PROs mentioned above for evaluating the quality of care and clinical performance.
Conclusion: Caregivers can certainly benefit from the NIH-provided measures, even if some of these sets are too narrow (such as those targeting only patients with sickle cell disease). These criteria are also more about quality of life, therefore this system of measures doesn’t allow to evaluate patients’ outcomes after treating particular diseases (for instance, ophthalmologic, cardiac or ENT ones).
The filtering system is quite simple, so providers won’t get lost in irrelevant measures. Moreover, HealthMeasures suggests a number of data collection tools appropriate either for research or clinical practice. This will definitely simplify the process of putting the measures to use.
As Mayo Clinic is well known for its 100+ years of experience combined with a focus on innovation and research, we have decided to review their approach to health outcomes too. They have defined a dense set of measures by combining the CMS suggestions and criteria of a high importance for their own practice:
- Mortality ratio (created from observed and expected mortality)
- Composite Measure for Patient Safety including the hospital rate of occurrence for the number of occasions (accidental puncture or laceration, postoperative hip fracture, death among surgical inpatients with treatable serious complications
- Readmission rates
- Transplant quality indicators (including heart, kidney, liver, lung and pancreas transplant)
Conclusion: This list of measures above would hardly satisfy the needs of each health system out there, since it doesn’t go that far from mortality-readmission rates. However, we don’t exclude the possibility that Mayo Clinic has more outcome measures that are just not in public access. Or, they may be in the middle of developing new quality sets for the future use.
A sneak peek at international practice: UK
The UK National Institute for Health and Care Excellence (NICE) is a national guide for providers. This organization develops quality standards and indicators to improve health and social care.
Currently, there are 127 condition and process standards supported by the sets of certain quality statements, where each statement is a separate measure. For example, if we take stroke in adults, there are 7 statements available, of which 4 are outcomes-focused:
- Prompt admission to specialist acute stroke units (target outcome: mortality rates of patients with stroke)
- Early supported discharge (target outcome: length of hospital stay)
- Return to work (target outcome: quality of life for post-stroke patients)
- Regular review of rehabilitation goals (target outcome: readmission rates of post-stroke patients)
Apart from listing the standards, NICE also offers guidelines on development sources, suggests tools for measurements, defines the terms used in quality statements as well as describes why each statement matters and how it impacts care quality.
Using the NICE’s quality standards, the Health and Social Care Information Centre (HSCIC) collaborates with such organizations as Healthcare Quality Improvement Partnership (HQIP), Royal College of Surgeons (RCS) and others to create national audits covering a range of chronic diseases and related conditions.
These audits include valuable findings for healthcare data analytics, supported by recommendations and overviews on a particular disease. For example, the national diabetes audit contains a table with a set of essential outcomes for diabetes patients – their HbA1c, cholesterol and blood pressure:
Conclusion: While not covering every disease, NICE tries to simplify caregivers’ life by offering all the available information on a silver platter, with comprehensive explanations re particular outcomes. The filtering system goes top down from a particular disease, allowing to see all current statements and standards in one place, not like in AHRQ’s NQMC, where caregivers need to think about an appropriate keyword, then check the ‘Outcome’ box and then manually sort the results to extract valuable measures.
How caregivers can harness health outcome analytics
Summing it all up, caregivers have a number of options to go for. They can stick to already available metrics, independently research in the field of future measures defined by CMS, or adopt the experience of fellow healthcare organizations nearby and around the globe.
When it comes to researching and developing measures for the purpose of internal healthcare data analytics, we suggest a more systematic solution. For this, a caregiver or an alliance of providers can elaborate on a consistent model of health outcome analytics. They can put particular departments at the core of the framework in question and determine certain sets of diseases and conditions appropriate for each department.
Also, providers can create different sets of the measures on surgical interventions, where some will be department-specific while others will relate to a particular disease or a condition.
Guided by such a system, caregivers will be able to see a full picture of their patients’ treatment outcomes with all ups and downs. Thus, a health organization’s performance will become easier to monitor, evaluate and correct.
By Lola Koktysh, Healthcare Industry Analyst at ScienceSoft