Health-Related Quality of Life in High-Cost, High-Need Populations

ABSTRACT

Objectives: To characterize patient-reported health and assess the psychometric performance of health-related quality of life (HRQOL) in high-cost, high-need (HCHN) populations.

Study Design: Retrospective longitudinal study examining health care utilization, expenditures, and patient-reported health comparing a baseline (year 1) and follow-up year (year 2).

Methods: The sample includes adults (n = 46,934) participating in the Medical Expenditure Panel Survey between 2011 and 2016. We estimated HRQOL for each sample member using the physical and mental health scales from the Medical Outcomes Study Short Form 12. We compared HRQOL stratified by HCHN, defined as patients whose baseline (year 1) demographics, utilization, and clinical characteristics predicted top decile health spending in year 2. Analyses assessed the validity, reliability, and responsiveness of physical and, separately, mental health scales.

Results: Among HCHN adults, the physical health scale exhibited robust measure validity, reliability, and responsiveness across all age groups; the mental health scale did not. Mean physical health was 1.25 SDs lower in HCHN vs other patients (37.9 vs 51.0 on a 0-100 scale increasing in self-perceived health; pooled SD, 10.5). Regressions indicated that a 0.5-SD increase in year 1 physical health among HCHN adults predicted a 5-percentage-point (10%) decrease in the probability of top decile health spending in year 2. In contrast to health care spending, HRQOL did not exhibit reversion to the mean in HCHN patients.

Conclusions: Patient-reported health outcomes remain poor in HCHN populations, even after health care utilization recedes. HRQOL is a promising outcome measure for HCHN-focused payment and delivery interventions.

Am J Manag Care. 2023;29(7):362-368. https://doi.org/10.37765/ajmc.2023.89396

Takeaway Points

We provide the first psychometric assessment of health-related quality of life (HRQOL) among high-cost, high-need (HCHN) populations. This is important for the following reasons:

• Health care utilization and expenditures revert to the mean among HCHN populations.
• In stark contrast, patient-reported health outcomes do not; poor self-rated health persists.
• Outcome measurement for HCHN-focused interventions must move beyond narrowly focused health care utilization and expenditures to appropriately capture patient suffering.
• We demonstrate that HRQOL is a psychometrically sound candidate for patient-reported outcome measurement in HCHN-focused interventions.

When states aim to improve care among Medicaid “superutilizers,”¹ federal alternative payment models incorporate refinements targeting high-cost, high-need (HCHN) populations,² and novel HCHN-focused quality improvement interventions are developed and tested by payers and health systems, the dominant measures of success are utilization and spending.³ Typical interventions augment standard care with enhanced care coordination and patient engagement strategies such as home visits, access to 24/7 nurse call lines, increased monitoring during transitions across care settings, and by connecting patients with social service providers.^4,5 Associated impact evaluations rely heavily on administrative data on resource utilization as key outcome measures.^6,7 Although elevated use of expensive health care services can be an independent marker of vulnerability in its own right,⁸ overreliance on utilization measures has several shortcomings. Common algorithms to identify HCHN patients for targeted intervention in health systems, due to reliance on health care utilization, can reinforce racial and other biases in access to care (eg, sex, geographic, economic).⁹ Furthermore, utilization measures ignore outcomes that are most valued by patients.

Patient-reported outcome measures offer one avenue to improve measurement of patient progress. These data capture patient perceptions of their underlying health status and functioning—fundamental dimensions to overall quality of life.¹⁰ Although patient-reported outcome measures have ample evidence of validity, reliability, and responsiveness in broad populations,^11,12 one may worry that for HCHN patients, such measures will suffer from floor effects (ie, everyone is in poor health) or may not add value over more widely available utilization measures such as emergency department (ED) visits and hospitalizations. In this article, we ask whether a well-validated patient-reported outcome measure construct—health-related quality of life (HRQOL)—identifies meaningful differences within HCHN patients and whether and how HRQOL changes over time for this population. In addition to characterizing levels and trends in HRQOL among HCHN patients, we test the scientific appropriateness of adopting HRQOL for population and performance monitoring. Our overarching goal is to understand whether the current utilization-centric focus on HCHN populations and initiatives that serve them could benefit from measuring and prioritizing patient-centric health outcome improvement.

METHODS

Data

We used the Medical Expenditure Panel Survey (MEPS) to characterize patterns of self-reported physical and mental health and to assess the performance of these measures among HCHN populations. The analytic sample of 46,934 adults 18 years and older drew from 5 two-year panels of the Agency for Healthcare Research and Quality’s nationally representative MEPS spanning 2011 to 2016. The MEPS includes patient-reported health measures and health care utilization and cost outcomes among HCHN populations.¹³ The eAppendix Figure (eAppendices available at ajmc.com) depicts the timing of data capture by round for each study measure described below.

Measures

HRQOL. HRQOL reflects an individual’s health and its impact on several domains, including physical, mental, emotional, and social functioning. A large literature validates the use of HRQOL across clinical and community-based settings (examples include Moriarty et al,¹⁴ Slabaugh et al,¹⁵ Zack,¹⁶ and Vilagut et al¹⁷).¹⁸ The widespread, standardized collection and dissemination of HRQOL data enable interested stakeholders to benchmark health across place, time, and population subgroups; this makes HRQOL an attractive measure in new settings and for new subgroups—such as HCHN populations—for whom no such benchmarks currently exist. In contrast to condition-specific patient-reported outcome measures, the global nature of HRQOL is potentially valuable in the context of HCHN populations, which are characterized by a high prevalence of multimorbidity.¹⁹ To date, HRQOL validation has not targeted HCHN populations.

The MEPS captures HRQOL using the Medical Outcomes Study Short Form 12 (SF-12, version 2), developed in the 1980s.²⁰ The SF-12 is constructed of 2 subscales, the Physical Health Composite Scale and Mental Health Composite Scale, referred to as “physical health” and “mental health” throughout this article. The physical and mental health composites use 12 identical items (eAppendix Table 1); however, the (proprietary) item weighting differs across them. Both composites are scaled to have a mean (SD) of 50 (10). Higher scores represent better health.²⁰ Numerous MEPS-specific studies have demonstrated the reliability and validity of both scales in cohorts defined by the presence of either specific symptoms (eg, a study of individuals with noncancer pain),²¹ or specific conditions (eg, a study of individuals with diabetes).²²

Other measures. We used the MEPS summary measures capturing annual total ED visits and overnight inpatient stays, as well as total annual health care expenditures. We adapted the method developed by Fleishman and Cohen²³ to identify the following chronic conditions: anxiety, arthritis, asthma, cancer, cerebrovascular disease, diabetes, emphysema, heart disease, high blood pressure, high cholesterol, mood disorder (including depression), schizophrenia, and substance use disorder. eAppendix A provides further detail.

We used the following measures as covariates in adjusted analyses, or to stratify analyses as appropriate: age, sex, education, marital status, income, and geographic region. MEPS participants who were pregnant were excluded from the analytic sample. In keeping with related work,²¹ we incorporated 2 common measures of self-reported health status that are conceptually related to HRQOL for validity testing: self-reported general health and self-reported mental health.

Empirical Approach

HCHN cohort definition. Following related literature^23,24 and common implementation practice,²⁵ we identified HCHN sample members using predicted spending. We estimated the probability that an individual is in the top expenditure decile next year (year 2) as a function of baseline (year 1) measures of health care utilization, expenditures, chronic conditions, and sociodemographics. We then selected sample members in the top decile of predicted year 2 spending for our analytic sample. eAppendix B includes methodological details and regression results.

Psychometric assessment. Following the CDC,²⁶ the scientific literature (eg, Aaronson et al²⁷), the National Quality Forum,²⁸ and CMS,²⁹ we designed empirical tests assessing the validity, reliability, and responsiveness of HRQOL among our HCHN cohort.

Validity. To assess validity, we asked the following questions:

1. Does HRQOL correlate with other well-established, conceptually related patient-reported outcome measures in a cross-section (construct validity)?
2. Is there appreciable regression to the mean in HRQOL, biasing longitudinal comparisons (internal validity)?
3. Does HRQOL predict future outcomes of interest (predictive validity)?

We assessed construct validity by estimating the cross-sectional association between year 1 values of mental health and physical health and the year 1 number of chronic conditions among HCHN sample members, adjusting for sociodemographic factors (age, sex, and education). This “sum-of-conditions” approach proxies the Charlson and Elixhauser indices, both validated comorbidity measures.³⁰ A negative correlation or gradient between the number of chronic conditions and HRQOL suggests that the latter demonstrates appropriate construct validity.

We estimated 1-year changes (year 2 minus year 1) to assess the extent of regression to the mean in HRQOL among HCHN sample members, a key internal validity check. Spending- and utilization-based measures exhibit considerable regression to the mean among HCHN patients.^5,31 To our knowledge, however, no analogous evidence tests whether the pattern holds true for HCHN patients’ HRQOL. We compared the pattern of mental health and physical health scales over time with that of other patient-reported outcomes—specifically, single-item self-reported general and mental health—as well as with those exhibited by traditional resource-based measures capturing elevated overall expenditures and ED and inpatient use. Additionally, we compared the longitudinal HRQOL patterns of HCHN vs non-HCHN patients, formalized using a difference-in-differences framework (associated model specification detailed in eAppendix C).

We estimated the capacity of HRQOL to predict future expenditures and utilization—indicators of HCHN persistence—as our key predictive validity test. To our knowledge, no research exists indicating whether HRQOL independently predicts expenditures and utilization among HCHN populations. Using logit models, we estimated whether an individual is in the top decile of medical expenditures in year 2 as a function of year 1 measures of HRQOL scales of mental and physical health among HCHN patients, controlling for year 1 sociodemographic characteristics (results in eAppendix Table 2). A positive coefficient exhibiting meaningful magnitude and statistical significance is indicative of predictive validity. We estimated similar models for year 2 top decile ED visits and for any inpatient hospitalization in year 2.

Reliability

Our primary reliability concern was ensuring HRQOL’s robustness across the variety of HCHN definitions employed by different stakeholders. The 2 main approaches to HCHN cohort definition are prospective assignment—the approach we used—which employs predictive modeling to identify those patients who are at elevated probability of HCHN status in the upcoming year, and rules-based assignment, employing cutoffs based on the prior year’s utilization and/or expenditures. The key related reliability concern for HRQOL is whether its values are comparable across the various HCHN cohort classifications and whether these values differ meaningfully relative to non-HCHN patients. We tested this by computing mental and physical health scales for our preferred cohort and 4 additional (rules-based) cohorts: top decile expenditures in year 1; 2 or more ED visits in year 1; any hospitalization in year 1; and 3 or more chronic conditions in year 1. We computed unadjusted means overall and stratified by age, comparing values within each of the 5 HCHN definitions and across the HCHN vs never-HCHN cohorts.

Responsiveness

We assessed the potential responsiveness of HRQOL by exploring whether HRQOL measures exhibit meaningful spread within the HCHN cohort. Spread can indicate unwarranted variation amenable to health care intervention.³² We were particularly mindful of the potential presence of floor effects among HCHN patients, given their high levels of morbidity. The extent of underlying spread also has implications for statistical power requirements, a key implementation consideration. We provided simple exploratory power calculations assuming effect sizes of 0.3, 0.5, and 1 SD. Values of 0.3 and 0.5 SD reflect “minimally important differences” in HRQOL documented in clinical studies of treatments targeting vulnerable populations (eg, Jayadevappa et al,³³ Norman et al³⁴) and in population-based studies (eg, Alghnam et al,³⁵ Baragaba et al,³⁶ Utah Department of Health³⁷). Finally, we estimated the marginal effect of increasing year 1 HRQOL by 0.3 and 0.5 SDs on the probability of year 2 top decile expenditure, using the associated predictive validity model discussed above. This model, although observational, demonstrates whether feasible improvements in mental and physical health might translate into future decreases in resource-intensive utilization.

RESULTS

Table 1 [part A and part B] displays descriptive characteristics for the entire sample and separately across HCHN and non-HCHN sample members (subsequently referred to as HCHNs and non-HCHNs for brevity). The sociodemographic composition across the 2 strata were fairly similar, with the notable exceptions of age and sex. The mean ages of HCHNs and non-HCHNs were 64.2 and 45.5 years, respectively. A higher percentage of HCHNs were women compared with non-HCHNs (58.2% vs 49.5%).

Our key measures of interest—HRQOL measures of mental and physical health—were lower in HCHNs than in non-HCHNs. However, the relative difference across the 2 measures was much larger for physical than for mental health. The mean physical health score of 37.9 among HCHNs was 1.25 SDs lower than the mean of 51.0 among non-HCHNs (using the full sample SD). In contrast, the mean mental health score for the 2 groups differed by only 3.4 points (48.6 vs 52.0 among HCHNs vs non-HCHNs), or a 0.35-SD difference (again, using the full sample SD).

Figure 1 and Figure 2 display cross-sectional evidence regarding reliability and concurrent validity. The panels comprising Figure 1 plot HRQOL distributions stratified by age and various definitions of HCHN status. The age stratification helped us rule out the possibility that the HCHN vs non-HCHN differences in HRQOL are driven solely by differences in age distribution. Computing mental and physical health across different HCHN definitions served as our primary reliability test, assessing the robustness of HRQOL’s use across varying but related HCHN cohorts. Within each age group, the never-HCHN sample—defined as not meeting any of the HCHN definitions—exhibited considerably higher physical health scores relative to all HCHN populations (all differences statistically significant). Moreover, the distributions exhibited considerable spread within HCHNs, providing reassurance regarding floor and ceiling effect concerns. The mental health gradient across HCHN and never-HCHN populations, in contrast, differed markedly across age strata. In younger age strata, mental health was much lower among HCHN than never-HCHN adults. For older age strata, which comprised the majority of HCHN adults, there was very little gradient in mental health. Small sample sizes for the younger age strata indicate a need for caution when interpreting these different age patterns; notably, only 393 of the HCHN cohort members were younger than 45 years (age strata sample sizes displayed in eAppendix Table 3).

Figure 2 demonstrates the results from our concurrent validity analysis, which assessed the association between year 1 HRQOL and chronic condition count among HCHN adults, adjusted for age, sex, and education. The estimates indicated that both mental and physical health declined fairly monotonically as the condition count increases, demonstrating their validity as benchmarked against increasing levels of comorbidity. Here again we saw that the gradient is stronger for physical than for mental health. For example, moving from 3 chronic conditions (the mean among HCHN adults) to 5 or more chronic conditions (22.4% of the HCHN sample) was associated with an almost 5-point decrease in physical health, compared with an approximately 3.5-point decrease in mental health.

Table 2 incorporates longitudinal analyses. Patterns for utilization- and resource-based measures among HCHNs comparing years 1 and 2 demonstrated mean reversion patterns comparable with those of the broader related literature.^5,31 For example, among people in the top decile of predicted year 2 spending, 83.7% had observed top decile spending in year 1, but only 47.6% remained in the top decile of observed spending in year 2. In stark contrast, neither the mental nor physical health scores among the HCHN cohort exhibited qualitatively meaningful or statistically significant year-over-year regression to the mean. The measures of patient-reported fair or poor health exhibited similar stability, providing additional evidence that decreases in resource-intensive utilization do not translate to comparable improvements in patient-reported outcome measures in an observational context. The differences-in-differences results, comparing year-over-year changes across HCHN and non-HCHN cohorts, demonstrated that although changes in expenditures and resource utilization among HCHN adults varied meaningfully from those among non-HCHN adults, changes in patient-reported outcome measures did not.

Estimates from the predictive validity regressions (Table 3) indicated that among the HCHN cohort, year 1 physical health exhibits robust, independent predictive capacity across all year 2 resource utilization measures (odds ratios in eAppendix Table 4). In contrast, year 1 mental health exhibited statistically significant predictive ability only for top decile expenditure membership. Moreover, the magnitude of the association between year 1 mental health and year 2 resource utilization measures was uniformly smaller than that of physical health. Table 3 displays predicted probabilities derived from the top decile expenditures regression. A 0.3-SD improvement in year 1 physical health was independently associated with a 3-percentage-point (6%) decrease in the probability of year 2 top decile expenditure membership. A 0.5-SD improvement in year 1 physical health was associated with a 5-percentage-point (10%) decrease, and a 1-SD improvement in year 1 physical health was associated with a 9-percentage-point (19%) decrease.

Power calculations indicated that for both the mental and physical health scores, fewer than 20 HCHNs are required in each treatment arm to achieve 80% power for expected impacts of 1 SD (eAppendix Table 5). Sample sizes of approximately 65 individuals per group were needed to achieve 80% power for expected impacts of 0.5 SD, and approximately 145 per group were needed for expected impacts of 0.3 SD. Assuming that 10% of the commonly cited primary care panel size of 2500 patients qualifies as HCHN suggests that initiatives at the practice level—even those using within-practice treatment and control arms—may be suitably powered for expected impacts of 0.5 SD or higher.³⁸ Expected impacts on the order of 0.3 SD would likely require cross-practice treatment and comparison arms. Populations tracked by multipractice medical groups, even modestly sized integrated health systems, payers, and policy makers, easily exceed all of the minimum sample size requirements.

DISCUSSION

We found that HCHN adults experienced lower HRQOL relative to non-HCHN adults and that this gradient was appreciably stronger for physical health compared with mental health. Together, our empirical tests suggested that physical health exhibited strong validity, reliability, and responsiveness as a measurement tool for HCHN adults, whereas mental health—which, unlike physical health, improves at older ages—was weaker along certain dimensions of validity and responsiveness. Unlike measures of resource utilization, HRQOL exhibited minimal regression to the mean among HCHN adults. Our findings imply that HRQOL represents an important, distinct construct from the utilization-based metrics currently used to hold providers or payers accountable for care of HCHN populations.

Limitations

Several limitations of the current study must be addressed prior to the use of HRQOL for higher-stakes HCHN measurement contexts, such as those involving meaningful financial or reputational impacts. First, our findings do not address the internal validity threat arising from likely differences in patient case mix, which would require risk adjustment. Moreover, tracing longer-term trajectories of the interplay between HRQOL and resource utilization among HCHNs could improve predictive validity and yield clues on effective intervention design. Further work should explore trade-offs associated with using a shorter HRQOL instrument among HCHNs—for example, a single self-rated health item—to balance psychometric soundness vs measurement burden, as was done for non-HCHN populations in the seminal HRQOL development work.³⁹

Our estimates also revealed a need for further exploration of the divergent mental health patterns across different HCHN age strata. Statistical power limitations precluded our ability to produce the associated age-stratified validity assessments. Exploring the potential for using mental health scores as measurement tools in interventions targeting younger ages is a potentially promising direction for future research. Additionally, further work on the interplay among sex, HCHN status, and HRQOL is important, as women seem to experience a higher prevalence of HCHN and have a different set of factors that drive HRQOL relative to men.⁴⁰

CONCLUSIONS

Our work contributes to the call for capturing the patient voice in measuring HCHN outcomes. We demonstrate that HCHN adults experienced significant health burdens that persist after their health care interactions recede. As such, an overemphasis on utilization when targeting care delivery efforts to HCHN adults excludes many who are still suffering and ignores the full burden of illness, which extends far beyond its cost.

Author Affiliations: Tuck School of Business, Dartmouth College (LJL), Hanover, NH; The Dartmouth Institute, Geisel School of Medicine, Dartmouth College (MT), Hanover, NH; Harvard School of Public Health, Harvard University (EM), Boston, MA.

Source of Funding: Agency for Healthcare Research and Quality’s Comparative Health System Performance Initiative under grant No. 1U19HS024075.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (LJL, EM); acquisition of data (LJL, MT); analysis and interpretation of data (LJL, MT); drafting of the manuscript (LJL); critical revision of the manuscript for important intellectual content (LJL, MT); statistical analysis (LJL, MT); obtaining funding (EM); administrative, technical, or logistic support (LJL, MT); and supervision (LJL, EM).

Address Correspondence to: Lindsey Jeanne Leininger, PhD, Tuck School of Business, Dartmouth College, 100 Tuck Hall, Hanover, NH 03755. Email: lindsey.j.leininger@tuck.dartmouth.edu.

REFERENCES

1. Medicaid Accelerated eXchange (MAX) series program: final report. New York State Department of Health. Updated December 2022. Accessed June 6, 2023. https://www.health.ny.gov/health_care/medicaid/redesign/dsrip/pps_workshops/max/docs/9-21_12-22_final_rpt.pdf

2. Colla CH, Lewis VA, Kao LS, O’Malley AJ, Chang CH, Fisher ES. Association between Medicare accountable care organization implementation and spending among clinically vulnerable beneficiaries. JAMA Intern Med. 2016;176(8):1167-1175. doi:10.1001/jamainternmed.2016.2827

3. Das LT, Abramson EL, Kaushal R. High-need, high-cost patients offer solutions for improving their care and reducing costs. NEJM Catalyst. Published online February 5, 2019. https://catalyst.nejm.org/high-need-high-cost-patients-solutions/

4. Long P, Abrams M, Milstein A, et al, eds. Effective Care for High-Need Patients: Opportunities for Improving Outcomes, Value, and Health. National Academy of Medicine; 2017. Accessed October 4, 2019. https://lccn.loc.gov/2017041343

5. Johnson TL, Rinehart DJ, Durfee J, et al. For many patients who use large amounts of health care services, the need is intense yet temporary. Health Aff (Millwood). 2015;34(8):1312-1319. doi:10.1377/hlthaff.2014.1186

6. Finkelstein A, Zhou A, Taubman S, Doyle J. Health care hotspotting — a randomized, controlled trial. N Engl J Med. 2020;382(2):152-162. doi:10.1056/NEJMsa1906848

7. Komaromy M, Bartlett J, Gonzales-van Horn SR, et al. A novel intervention for high-need, high-cost Medicaid patients: a study of ECHO care. J Gen Intern Med. 2020;35(1):21-27. doi:10.1007/s11606-019-05206-0

8. Ryan J, Abrams MK, Doty MM, Shah T, Schneider EC. How high-need patients experience health care in the United States. The Commonwealth Fund. December 9, 2016. Accessed June 6, 2023. https://bit.ly/44pWWC3

9. Obermeyer Z, Powers B, Vogeli C, Mullainathan S. Dissecting racial bias in an algorithm used to manage the health of populations. Science. 2019;366(6464):447-453. doi:10.1126/science.aax2342

10. Health-related quality of life (HRQOL): well-being concepts. CDC. Updated October 31, 2018. Accessed June 6, 2023.https://www.cdc.gov/hrqol/wellbeing.htm

11. Cella D, Hahn EA, Jensen SE, et al. Patient-Reported Outcomes in Performance Measurement. RTI Press; 2015. doi:10.3768/rtipress.2015.bk.0014.1509

12. Kallenberg GA. The origins and promise of PROMIS – Patient Reported Outcomes Measurement Information System. Workshop on Practice Based Research. December 5, 2015. Accessed October 4, 2019.
https://medschool.ucsd.edu/som/fmph/research/centers/integrativehealth/TrainingEducation/Documents/Kallenberg_The_Promise_of_PROMIS_11-27-15.pdf

13. Burns M, Mullahy J. Healthy-time measures of health outcomes and healthcare quality. National Bureau of Economic Research. August 2016. doi:10.3386/w22562

14. Moriarty DG, Zack MM, Kobau R. The Centers for Disease Control and Prevention’s healthy days measures – population tracking of perceived physical and mental health over time. Health Qual Life Outcomes. 2003;1:37. doi:10.1186/1477-7525-1-37

15. Slabaugh SL, Shah M, Zack M, et al. Leveraging health-related quality of life in population health management: the case for healthy days. Popul Health Manag. 2017;20(1):13-22. doi:10.1089/pop.2015.0162

16. Zack MM; CDC. Health-related quality of life – United States, 2006 and 2010. MMWR Suppl. 2013;62(3):105-111.

17. Vilagut G, Forero CG, Pinto-Meza A, et al. The mental component of the Short-Form 12 Health Survey (SF-12) as a measure of depressive disorders in the general population: results with three alternative scoring methods. Value Health. 2013;16(4):564-573. doi:10.1016/j.jval.2013.01.006

18. Search string “health-related quality of life.” ClinicalTrials.gov. https://clinicaltrials.gov/ct2/results?cond=&term=health-related+quality+of+life&cntry=&state=&city=&dist=

19. Mann C. Targeting Medicaid super-utilizers to decrease costs and improve quality. Medicaid. 2013. Accessed October 30, 2019. https://www.medicaid.gov/federal-policy-guidance/downloads/cib-07-24-2013.pdf

20. Hays RD, Sherbourne CD, Mazel R. User’s Manual for the Medical Outcomes Study (MOS) Core Measures of Health-Related Quality of Life. RAND Corporation; 1995.

21. Hayes CJ, Bhandari NR, Kathe N, Payakachat N. Reliability and validity of the Medical Outcomes Study Short Form-12 version 2 (SF-12v2) in adults with non-cancer pain. Healthcare (Basel). 2017;5(2):22. doi:10.3390/healthcare5020022

22. Kathe N, Hayes CJ, Bhandari NR, Payakachat N. Assessment of reliability and validity of SF-12v2 among a diabetic population. Value Health. 2018;21(4):432-440. doi:10.1016/j.jval.2017.09.007

23. Fleishman JA, Cohen JW. Using information on clinical conditions to predict high-cost patients. Health Serv Res. 2010;45(2):532-552. doi:10.1111/j.1475-6773.2009.01080.x

24. Wherry LR, Burns ME, Leininger LJ. Using self-reported health measures to predict high-need cases among Medicaid-eligible adults. Health Serv Res. 2014;49(S2):2147-2172. doi:10.1111/1475-6773.12222

25. Jean-Baptiste D, O’Malley A, Shah T. Population segmentation and targeting of health care resources: findings from a literature review. Mathematica. December 1, 2017. Accessed October 4, 2019. https://www.mathematica.org/our-publications-and-findings/publications/population-segmentation-and-targeting-of-health-care-resources-findings-from-a-literature-review

26. Measuring Healthy Days: Population Assessment of Health-Related Quality of Life. CDC; 2000. Accessed October 4, 2019. https://www.cdc.gov/hrqol/pdfs/mhd.pdf

27. Aaronson N, Alonso J, Burnam A, et al. Assessing health status and quality-of-life instruments: attributes and review criteria. Qual Life Res. 2002;11(3):193-205. doi:10.1023/A:1015291021312

28. Patient-reported outcomes in performance measurement. National Quality Forum. December 2012. Accessed October 8, 2019. https://www.qualityforum.org/Publications/2012/12/Patient-Reported_Outcomes_in_Performance_Measurement.aspx

29. CMS Measures Management System Blueprint quickstart guide. CMS. December 2022. Accessed June 6, 2023. https://mmshub.cms.gov/sites/default/files/Quickstart.pdf

30. Austin SR, Wong Y-N, Uzzo RG, Beck JR, Egleston BL. Why summary comorbidity measures such as the Charlson Comorbidity Index and Elixhauser Score work. Med Care. 2015;53(9):e65-e72. doi:10.1097/MLR.0b013e318297429c

31. Chakravarty S, Cantor JC. Informing the design and evaluation of superuser care management initiatives: accounting for regression-to-the-mean. Med Care. 2016;54(9):860-867. doi:10.1097/MLR.0000000000000568

32. Wennberg JE. Tracking Medicine. Oxford University Press; 2010.

33. Jayadevappa R, Cook R, Chhatre S. Minimal important difference to infer changes in health-related quality of life—a systematic review. J Clin Epidemiol. 2017;89:188-198. doi:10.1016/j.jclinepi.2017.06.009

34. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003;41(5):582-592. doi:10.1097/01.MLR.0000062554.74615.4C

35. Alghnam S, Schneider EB, Castillo RC. Insurance status and health-related quality-of-life disparities after trauma: results from a nationally representative survey in the US. Qual Life Res. 2016;25(4):987-995. doi:10.1007/s11136-015-1126-0

36. Baragaba B, Alghnam S, Bernacki EJ. Work-related injuries and health-related quality of life among US workers: a longitudinal study of a population-based sample. J Occup Environ Med. 2016;58(4):385-390. doi:10.1097/JOM.0000000000000667

37. Health status in Utah: the Medical Outcomes Study SF-12. Utah Department of Health. 2001. Accessed June 6, 2023. https://ibis.health.utah.gov/ibisph-view/pdf/opha/publication/uhas/2001HSS_SF12.pdf

38. Yarnall KSH, Pollak KI, Østbye T, Krause KM, Michener JL. Primary care: is there enough time for prevention? Am J Public Health. 2003;93(4):635-641. doi:10.2105/ajph.93.4.635

39. Manning WG, Newhouse JP, Ware JE. The status of health in demand estimation; or, beyond excellent, good, fair, poor. In: Fuchs VR, ed. Economic Aspects of Health. University of Chicago Press; 1982:141-184.

40. Wlodarczyk D, Zietalewicz U. How gender-specific are predictors of post-MI HRQoL? a longitudinal study. Health Qual Life Outcomes. 2020;18(1):202. doi:10.1186/s12955-020-01439-4