Association Between Documented Severe Pain and Cognitive Impairment in Home Health Care Patients: Results from the National Outcome and Assessment Information Set Data

Abstract

Background:

Despite the growing importance of home health care (HHC) in the care of older adults with cognitive impairment, limited evidence exists about factors associated with documented severe pain among older adults receiving HHC.

Methods:

This secondary data analysis used a 5% random national sample of the 2017 national Outcome and Assessment Information Set (OASIS) data. Multivariable Poisson regression model was used to examine the association between documented severe pain, cognitive impairment, and a range of sociodemographic, clinical, and cognitive factors.

Results:

HHC patients (n = 183,038) were mean age 79.7 years, 61.7% female, and 78.6% non-Hispanic White. In multivariable models, cognitive impairment was associated with lower likelihood of documented severe pain (prevalence ratio [PR] = 0.96, confidence interval [CI] = [0.94–0.98]). Other factors independently associated with less documented severe pain included Hispanic ethnicity (PR = 0.92, CI = [0.89–0.95]), the oldest (≥85 years) groups (PR = 0.65, CI = [0.63–0.66]), male patients (PR = 0.83, CI = [0.82–0.85]), those with a formal diagnosis of Alzheimer's Disease Related Dementias (PR = 0.67, CI = [0.65–0.69]), and patients with verbal- and speech-related difficulty (PR = 0.83, CI = [0.80–0.86]). Patients with history of falls (PR = 1.18, CI = [1.16–1.20]), positive screen on the Patient Health Questionnaire-2 (PR = 1.36, CI = [1.31–1.41]), shortness of breath (PR = 1.14, CI = [1.12–1.14]), anxiety daily (PR = 1.16, CI = [1.14–1.18]), and anxiety daily or more often (PR = 1.40, CI = [1.37–1.43]) were more likely to have documented severe pain.

Conclusions:

HHC patients with cognitive impairment were less likely to have documented severe pain even with a range of sociodemographic, clinical, functional, and cognitive characteristics were considered. These findings may reflect a link between cognitive impairment missed opportunities for clinicians to provide pain management. Tailored interventions are needed to better assess and manage pain in this vulnerable group of HHC patients.

Key Message

We examined the association between cognitive impairment and documented severe pain in a national sample of home health care patients. Patients with cognitive impairment were less likely to have documented severe pain even when a range of sociodemographic, clinical, and cognitive characteristics were considered.

Introduction

Pain is an important symptom for the >3 million older Medicare beneficiaries who rely on home health care (HHC) for skilled nursing, therapy, social worker and aide services to remain living in their own homes.^1,2 Pain is associated with adverse patient outcomes, including unplanned hospitalizations,³ falls,⁴ and mortality.⁵ In the HHC setting, pain substantially limits the ability to participate in rehabilitative or physical therapy older adults ability to participate in rehabilitative therapy.⁶ The relationship between cognition and pain is important because although patient self-report has widely been accepted as gold standard for pain assessment,⁷ not all patients are able to communicate their pain experience. This is particularly important to understand in the HHC population given the high rates of cognitive impairment among those who utilize HHC.⁸ Data from the national field test of candidate standardized assessment data elements show that a substantial number of patients (31%) receiving HHC are unable to self-report for pain.⁹

Yet much of the literature about pain among older adults, including the literature about the importance of pain among older adults with cognitive impairment who may be unable to communicate about pain,^9,10 has focused on nursing homes. The single-site HHC agency samples^3,11–13 and small sample sizes^3,12 have limited the generalizability of the few studies that have focused on pain in the HHC population. The handful of studies that exist estimate rates of severe pain (i.e., pain that interferes with movement or activity daily or all of the time) in older adults receiving HHC in the range of 21.8% (n = 1340) to 59% (n = 87,000).^3,11,14 One single-site study that focused specifically on severe pain in HHC setting found that HHC patients with Alzheimer's disease and related dementias (ADRD) were less likely to have recorded severe pain than those without ADRD; furthermore, having severe pain significantly increased the risk of having unplanned admission to the hospital, and nursing home admission.³

In a similar study using Outcome and Assessment Information Set (OASIS) data from a single agency, researchers examined the relationship between ADRD diagnosis and analgesic use among HHC patients with daily pain and found that HHC patients with ADRD reporting daily pain were less likely to have any analgesic.¹² No study investigating pain using HHC sample^11,12,15 examined the influence of a patient's ability to self-report on pain experiences.

More information about pain in the HHC setting is needed to support the development of clinical guidelines and interventions that are tailored to the needs of HHC patients, particularly those with cognitive impairment who are unable to communicate their experience with pain. The goal of this study was to examine factors associated with severe pain, with a focus on factors related to cognitive impairment among older adults receiving HHC.

Methods

Study sample and data sources

This cross-sectional, descriptive study is a secondary data analysis of 5% random national sample of HHC recipients, ≥65 years who had a start-of-care (SOC) OASIS (OASIS, version C) in 2017. OASIS is a Centers for Medicare and Medicaid Services (CMS)-mandated patient assessment tool for collecting administrative and clinical data on all patients receiving HHC from Medicare-certified HHC agencies. HHC clinicians (Registered Nurses or Therapists) use OASIS to assess nearly 100 items, including clinical status, caregiver characteristics, and functional status.¹⁶ HHC clinicians collect OASIS data using a combination of assessment strategies, including observation, patient and/or caregiver interviews, and review of clinical documentation (e.g., hospital discharge summaries). The reliability of OASIS items ranges from fair to excellent.^17,18 The reliability of pain items on the OASIS is adequate (Kappa 0.60).¹⁹ Items about cognitive impairment have strong inter-rater reliability and gold standard validity.^20,21 In this study, the OASIS dataset was linked to Medicare Master Beneficiary Summary File (MBSF) to obtain sociodemographic data. Adelphi University Institutional Review Board approved the study protocol. A Data Use Agreement was reviewed and approved by CMS.

Measures

Outcome variable

The outcome of interest was severe pain defined using the M1240 on the SOC assessment. The OASIS item M1240 asks, “has this patient had a formal pain assessment using a standardized pain assessment tool (appropriate to the patient's ability to communicate the severity of pain)?” on a 3-item scale (0–2). We categorized those rated “1” (Yes, and it does not indicate severe pain), as having no identified severe pain. We categorized those rated “2” (“Yes, and it indicates severe pain.”) as having documented severe pain.

Key predictor

We identified HHC patients with cognitive impairment using the OASIS item M1700, which requires HHC clinicians to rate patient cognitive function on a five-item scale (0–4) with a score of “0” indicative of no cognitive impairment, and “4” severe cognitive impairment. Following the method of Burgdorf, Amjad, and Knowles, we categorized those rated “1,” “2,” “3,” or “4” as having identified cognitive impairment.²²

Covariates

Our analyses were adjusted for sociodemographic, clinical, and cognitive characteristics potentially associated with the report and management of pain.^3,12,15,23 Sociodemographic characteristics were obtained from the Medicare MBSF: age (65–74, 75–84, and ≥85 years), gender (male/female), dual Medicare and/or Medicaid eligibility status. Race/ethnicity was categorized as non-Hispanic White, non-Hispanic Black, Hispanic, and other race/ethnicity based on the Research Triangle Institute algorithm in the MBSF.

Clinical characteristics obtained from the OASIS included (1) comorbidities; (2) symptom burden: anxiety, presence of depression, and dyspnea; (3) functional status: Activities of Daily Living (ADL) dependency; (4) unmet caregiving needs; (5) prior inpatient stay (nursing home/hospital); and (6) hospitalization risk factors: previous 12-month fall history. The patient's comorbidity burden was assessed using the Charlson Comorbidity Index.²⁴ Dichotomous variables were created to identify frequency of anxiety (M1720) (none/less than daily/daily or more often), depressive symptoms (Patient Health Questionnaire-2 [PHQ-2] ≥3), and shortness of breath. To assess functional status, ADLs were categorized into four levels: (1) independent in all functions; (2) required assistance in one to three ADL functions (mild); (3) required assistance in four to six ADL functions (moderate); and (4) required assistance in seven and more ADL functions (severe), indicating complete dependence on someone to perform ADL functions.

To assess unmet caregiving needs, we generated a composite variable using M2102 in OASIS-C that asks the ability and willingness of nonagency caregivers to provide assistance in seven tasks, with responses that characterize the presence of an informal caregiver to provide assistance on a scale of 0 to 4. All seven tasks were recategorized into two levels to identify the unmet caregiving needs. The composite score for the seven tasks was categorized as: (1) 0 = no identified caregiving need; (2) caregiver unavailable in one to two tasks; (3) caregiver unavailable in three or more tasks. Finally, hospitalization risk factors included previous 12-month fall history, and prior stay (hospital, nursing home, or community [M1000]).

Cognitive characteristics included understanding of verbal content, speech and oral ability, impaired decision making, presence of disruptive behavior, and formal ADRD diagnosis. To measure ability to understand verbal content, we dichotomized the OASIS item M1220 (understanding of verbal content in patients own language [0–4]). We categorized those rated “0,” which corresponds to “understands: clear comprehension without cues or repetition,” and rated as “1” “usually understands: understands most conversations, but misses some part/intent of message”—as having ability to understand verbal content. We defined ratings of “2” and “3” as inability to understand verbal content. We used the OASIS item, M1230 (Speech and Oral) to identify patients unable to self-report. M1230 assesses speech (0 = Expressed complex ideas and 1 = minimal difficulty) (2 = expresses simple ideas with moderate difficulty, 3 = has severe difficulty, 4 = unable to express, 5 = unresponsive—has difficulty expressing basic ideas or needs).

To identify patients unable to self-report for pain, we used OASIS item M1230: Speech and Oral to measure ability to self-report. M1230 requires the HHC clinician to rate patients' ability to express themselves on a six-item scale. We categorized those rated as “1,” “2,” “3,” or “4” on the five-item scale, which ranges from ability to “express simple ideas” with “minimal difficulty” to “unresponsive” as being unable to self-report. We also dichotomized disruptive behavior (M1745) (once a month or more). HHC patients with a diagnosis of ADRD were identified from the OASIS using the list of ICD-10 codes based on the Chronic Condition Warehouse definition for ADRD.²⁵

Statistical analyses

We summarized patient characteristics, overall and stratified by cognitive impairment, using means and standard deviations (SDs) for continuous variables and frequencies and percentages for categorical variables. Chi-square tests, t-tests, and Wilcoxon rank-sum tests were used as appropriate to compare differences in characteristics by level of cognitive impairment. Due to the large sample sizes and to help identify clinical significance, we considered differences in percentages of ≥5% as important. We identify factors associated with severe pain, using a robust multivariable Poisson regression. The χ² goodness-of-fit test demonstrated that the data fit a Poisson distribution. Multicollinearity was assessed using variance inflation factor (VIF) using Fisher's scoring and Hessian weight,^26,27 and ADL levels were removed from the model as it was collinear with cognitive impairment (VIF = 25.82). Variables with VIF greater than 5 were removed.²⁸ All analyses were completed using SAS, version 9.4 (SAS Institute, Inc.). All statistical tests were two-sided and statistical significance was set at p < 0.05.

Results

Sample description

Our final sample for this analysis included 87,113 patients with cognitive impairment and 95,925 without cognitive impairment. The prevalence of severe pain overall was 23.4%. The group with cognitive impairment was older, 81.7 years old (SD: 8.25) versus 78.0 years old (SD: 8.2), and the group with identified cognitive impairment had a higher proportion of patients >85 years old (38.5%) compared with those without identified cognitive impairment (22.4%).

As compared with those without cognitive impairment, those with cognitive impairment were more likely to experience shortness of breath (73,500 [84.4%] vs. 69,562 [72.5%] patients), anxiety daily or more (23,541 [27.0%] vs. 10,762 [11.2%] patients), and depression (4271 [4.9%] vs. 2232 [2.3%] patients). Additionally, patients with cognitive impairment were more likely to experience severe ADL dependency (83,954 [96.4%] vs. 85,202 [88.8%] patients), be admitted to HHC from the community—without a prior inpatient stay within 14 days (31,262 [35.9%] vs. 23,525 [24.5%] patients), and have a history of falls (35,337 [40.6%] vs. 26,368 [27.5%] patients) compared with those without cognitive impairment. No significant differences were observed in the level of unmet caregiving needs (Table 1).

Table 1.

Sample Characteristics of Home Health Care Patients by Identified Cognitive Impairment

	Overall	No cognitive impairment	Cognitive impairment
	N = 183,038	n = 95,925 (52.4%)	n = 87,113 (47.6%)
Sociodemographic characteristics
Age (mean ± SD)	79.7 (8.53)	78.0 (8.17)	81.7 (8.50)
Age
65–74	61,473 (33.6%)	39,673 (41.4%)	21,800 (25.0%)
75–84	66,588 (36.4%)	34,774 (36.3%)	31,814 (36.5%)
≥85	54,977 (30.0%)	21,478 (22.4%)	33,499 (38.5%)
Sex
Female	112,944 (61.7%)	58,907 (61.4%)	54,037 (62.0%)
Male	70,094 (38.3%)	37,018 (38.6%)	33,076 (38.0%)
Race/ethnicity
Non-Hispanic White	143,814 (78.6%)	76,528 (79.8%)	67,286 (77.2%)
Non-Hispanic Black	19,853 (10.9%)	10,124 (10.6%)	9729 (11.2%)
Hispanic	12,212 (6.7%)	5712 (6.0%)	6500 (3.6%)
Other	7159 (3.9%)	3561 (3.7%)	3598 (4.1%)
Socioeconomic status
Type of insurance
Dual eligibility	2054 (1.1%)	842 (0.9%)	1212 (1.4%)
Medicaid FFS	934 (0.5%)	5482 (0.5%)	452 (0.5%)
Medicare FFS	127,037 (69.4%)	64,871 (67.6%)	62,166 (71.4%)
Any managed care	53,013 (29.0%)	29,730 (31.0%)	23,283 (26.7%)
Clinical characteristics
Shortness of breath	143,062 (78.2%)	69,562 (72.5%)	73,500 (84.4%)
Frequency of anxiety
None	97,022 (53.0%)	65,599 (68.4%)	31,423 (36.1%)
Less than daily	51,713 (28.3%)	19,564 (20.4%)	32,149 (36.9%)
Daily of more often	34,303 (18.7%)	10,762 (11.2%)	23,541 (27.0%)
Depression (PHQ-2 ≥ 3)
Yes	6503 (3.6%)	2232 (2.3%)	4271 (4.9%)
No	176,535 (96.5%)	93,693 (97.67%)	82,842 (95.1%)
Pain
Severe	42,833 (23.4%)	23,398 (24.4%)	19,435 (22.3%)
Nonsevere	140,205 (76.6%)	72,527 (75.6%)	67,678 (77.7%)
Charlson Comorbidity Index (mean ± SD)	0.39 (0.92)	0.38 (0.94)	0.39 (0.90)
Hospitalization risk
Falls	61,705 (33.7%)	26,368 (27.5%)	35,337 (40.6%)
Multiple hospitalizations (two or more) in the past six months	42,370 (23.2%)	20,132 (21.0%)	22,238 (25.5%)
ADL dependency
Independent	883 (0.5%)	8784 (0.8%)	135 (0.2%)
Mild dependence	4719 (2.6%)	3733 (3.9%)	986 (1.1%)
Moderate dependence	8280 (4.5%)	6242 (6.5%)	2038 (2.3%)
Severe dependence	169,038 (92.4%)	85,202 (88.8%)	83,954 (96.4%)
Caregiving needs
0 domain	174,353 (95.3%)	91,334 (95.2%)	83,019 (95.3%)
1–2 domains	5623 (3.1%)	2992 (3.1%)	2631 (3.0%)
3 or more domains	3062 (1.7%)	1599 (1.7%)	1463 (1.7%)
Prior inpatient stay
Hospital	88,397 (48.3%)	52,680 (54.9%)	35,717 (41.0%)
Nursing home	32,197 (17.6%)	15,955 (16.6%)	16,242 (18.6%)
None (Community)	54,787 (29.9%)	23,525 (24.5%)	31,262 (35.9%)
Cognitive characteristics
Understanding of verbal content
Yes	103,699 (56.7%)	83,040 (86.6%)	20,659 (23.7%)
No	79,339 (43.4%)	12,885 (13.4%)	66,454 (76.3%)
Speech and oral problem
Yes	15,553 (8.5%)	815 (0.9%)	14,738 (16.9%)
No	167,485 (91.5%)	95,110 (99.2%)	72,675 (83.1%)
Impaired decision making
Yes	38,447 (21.0%)	3709 (3.9%)	34,738 (39.9%)
No	144,591 (79.0%)	92,216 (96.1%)	52,375 (6.1%)
Any disruptive behavior
None	145,593 (79.5%)	91,717 (95.6%)	53,876 (61.9%)
Once a month or more	37,445 (20.5%)	4208 (4.4%)	33,237 (38.2%)
Diagnosis of ADRD	25,346 (13.9%)	3399 (3.5%)	21,947 (25.2%)

ADL, activities of daily living; ADRD, Alzheimer's disease and related dementias; FFS, fee-for-service; HHC, home health care; PHQ, Patient Health Questionnaire; SD, standard deviation.

Patients with cognitive impairment were more likely to have speech and swallow problems than patients without identified cognitive impairment (14,738 [16.9%] vs. 815 [0.9%] patients). They were also more likely to have a diagnosis of ADRD (21,947 [25.2%] vs. 3399 [3.5%] patients). Compared with patients without cognitive impairment, those with cognitive impairment were more likely to experience disruptive behavior at least monthly (33,237 [38.2%] vs. 4208 [4.4%] patients) and have impaired decision making (34,738 [39.9%] vs. 3709 [3.9%] patients).

Factors associated with documentation of severe pain

Table 2 shows patient characteristics associated with documented severe pain. In univariate analysis, cognitive impairment was associated with lower likelihood of documented severe pain (prevalence ratio [PR] 0.91, confidence interval [CI] = 0.90–0.93). This association persisted in the multivariable analysis; patients with cognitive impairment had lower rate of documented severe pain (PR = 0.96, CI = [0.94–0.98]) compared with those without cognitive impairment.

Table 2.

Association Between Home Health Care Patient Characteristics and Severe Pain

	PR (95% CI)	PR (95% CI)
Sociodemographic characteristics
Age
65–74		Ref.
75–84		0.79 (0.78–0.81)^***
≥85		0.65 (0.63–0.66)^***
Sex
Female		Ref.
Male		0.83 (0.82–0.85)^***
Race/ethnicity
Non-Hispanic White		Ref.
Non-Hispanic Black		1.03 (1.00–1.05)
Hispanic		0.92 (0.89–0.95)^***
Other		0.98 (0.94–1.02)
Socioeconomic status
Type of insurance
Dual eligibility		1.12 (1.05–1.20)^***
Medicaid FFS		0.94 (0.84–1.06)
Medicare FFS		0.98 (0.94–1.02)
Any HMO		Ref.
Clinical characteristics
Shortness of breath		1.14 (1.12–1.16)^***
Frequency of anxiety
None		Ref.
Less than daily		1.16 (1.14–1.19)^***
Daily or more often		1.40 (1.37–1.43)^***
Depression (PHQ-2 ≥ 3)
Yes		1.36 (1.31–1.41)^***
No		Ref.
Charlson Comorbidity Index (mean ± SD)		0.92 (0.91–0.93)^***
Hospitalization risk
Falls		1.19 (1.17–1.21)^***
Caregiving assistance needed but unavailable
0 domain		Ref.
1–2 domains		0.98 (0.94–1.03)
3 or more domains		1.04 (0.98–1.10)
Prior inpatient stay
Hospital		1.12 (1.09–1.16)^***
Nursing home		1.00 (0.96–1.03)
None (Community)		1.21 (1.17–1.25)^***
Cognitive characteristics
Cognitive impairment
None		Ref.
Yes	0.91 (0.90–0.93)^***	0.96 (0.94–0.98)^***
Understanding of verbal content
Yes		1.03 (1.01–1.06)^**
No		Ref.
Speech and oral problem
Yes		0.83 (0.80–0.86)^***
No		Ref.
Impaired decision making
Yes		1.01 (0.98–1.04)
No		Ref.
Any disruptive behavior
None		Ref.
Once a month or more		1.01 (0.98–1.04)
Diagnosis of ADRD		0.67 (0.65–0.69)^***

Formal pain assessment (M1240) was analyzed using robust Poisson regression. The result is presented as PR with 95% CI.

p < 0.01.

***

p < 0.001.

CI, confidence interval; HMO, health maintenance organization; PR, prevalence ratio.

Several other characteristics were also associated with documented severe pain in this multivariable model. Among sociodemographic characteristics, older age was associated with a lower likelihood of having severe pain documented. Compared with residents 65–74 years of age, prevalence of documented severe pain was lower in groups 75 to 84 years of age (PR = 0.79, CI = [0.78–0.81]) and those 85 years of age or older (PR = 0.65, CI = [0.63–0.66]). Being male was associated with a lower likelihood of documented severe pain (PR = 0.83, CI = [0.82–0.85]) compared with female. Prevalence of documented severe pain was lower among Hispanic patients (PR = 0.92, CI = [0.89–0.95]) compared with non-Hispanic Whites.

Among clinical and cognitive characteristics, prevalence of documented severe pain was lower among patients with a diagnosis of ADRD (vs. no ADRD) (PR = 0.67, CI = [0.65–0.70]), patients with cognitive impairment (PR = 0.96, CI = [0.94–0.98]) compared with those without cognitive impairment. The prevalence of documented severe pain was lower among patients with verbal- and speech-related difficulty (PR = 0.83, CI = [0.80–0.86]) compared with those without speech and oral problem. In terms of symptoms, the presence of depression indicated by a positive screen on the PHQ-2 (PHQ-2 score ≥3) (PR = 1.36, CI = [1.31–1.41]), shortness of breath (PR = 1.14, CI = [1.12–1.16]), anxiety daily (PR = 1.16, CI = [1.14–1.19]), and daily or more often (PR = 1.40, CI = [1.37–1.43]) were more likely associated with severe pain documented.

Finally, prevalence of documented severe pain was greater among patients identified at higher risk for hospitalization due to a history of falls in the past 12 months (PR = 1.18, CI = [1.17–1.21]), patients with an inpatient stay in a hospital within 14 days before admission to HHC (PR = 1.12, CI = [1.09–1.16]), and those admitted directly from the community (PR = 1.21, CI = [1.17–1.25]).

Discussion

In this study, we examined the association of cognitive impairment with documentation of severe pain among older adults receiving HHC using nationwide OASIS data. We found that cognitive impairment was associated with lower likelihood of documented severe pain; the association persisted in multivariable models that considered a wide variety of sociodemographic, clinical, and cognitive factors. These findings raise important concerns about inadequacies in pain identification and management for vulnerable subpopulations of older Americans in the HHC setting. This study expands on similar findings from prior studies of particularly in the nursing home setting and extends findings from prior single-site studies.^12,15

Our findings about the association between documented severe pain and cognitive impairment are largely consistent with those of previous studies in the HHC population^3,12 and nursing home or hospital setting.^29–32 This finding is consistent with previous research, which has shown that pain is typically underreported for more cognitively impaired patients compared with patient without cognitive impairment.^3,15,33,34 Given the important relationship between cognitive impairment and documented severe pain, the finding that inability to self-report pain was associated with less documented severe pain is particularly important. Self-report of pain is generally considered the gold standard of pain management.^35,36 Inability to self-report is a risk for suboptimal pain management. For patients with cognitive impairment, who are unable to self-report, pain assessment strategies must rely on the ability of the HHC nurse or therapist to infer pain using direct observation of nonverbal behaviors^37–39 such as grimacing, guarding; such assessment can be difficult for clinicians.⁴⁰ The OASIS item M2401 intervention synopsis, option “d” focuses on intervention(s) to monitor and mitigate pain.

This response option emphasizes the need to use standardized, validated pain assessment conducted to determine that a patient has no pain. However, the current OASIS does not include a pain scale for assessment of patients with cognitive impairment.

There remains an opportunity to improve assessment of pain for patients with cognitive impairment in the HHC setting. Valid pain assessment scales for patients with cognitive impairment or otherwise unable to self-report pain, such as the Pain Assessment in Advanced Dementia (PAINAD), which assesses breathing, vocalizations, facial expression, body language, and consolability are recommended.^37,38 The absence of a valid scale within the OASIS to support clinician assessment of pain within the HHC setting is a gap that may be amenable to interventions, including ongoing education of HHC nurses and therapists about pain assessment for patients with cognitive impairment and adoption of HHC agency-specific best practices specific to pain assessment for patients with cognitive impairment.

In addition, other patient characteristics apart from cognitive impairment that were significantly associated with documented severe pain in multivariable models are also aligned with findings from the nursing home or hospital setting. Most notably, other analyses using multivariable models of nursing home population have found that racial/ethnic minority (Black and Hispanic Americans) nursing home residents were at increased risk of under-reporting of pain than their non-Hispanic White counterparts.^32,41 Similarly, prior research on pain in the nursing home population has shown that minoritized populations were less likely to have any pain documented compared with non-Hispanic White Americans.^30,32,42 Of note, in our study, Black race was not significant. The disparity we observed is likely due to a combination of underassessment and language discordance,⁴³ including under-reported pain severity because of inherent biases.⁴⁴ HHC agencies serving predominantly Hispanic population may need to establish agency-level strategies to monitor the quality of culturally sensitive communication between Hispanic patients and HHC clinicians in efforts to reduce disparities in pain recognition and management.

HHC agencies serving predominantly Hispanic population may need to establish agency-level strategies to monitor the quality of culturally sensitive communication between Hispanic patients and HHC clinicians in efforts to reduce disparities in pain recognition and management.

The HHC setting, compared with inpatient settings such as hospital or nursing home present unique challenges for clinicians in the assessment and management of pain. For example, in the nursing home setting, the MDS 3.0 includes four behavioral pain indicators, and staff are guided to document if these signs or behaviors are observed or reported during the five-day look-back period.⁴⁵ Such careful pain assessment is not yet the standard in HHC. However, national efforts focused on standardizing data assessment elements across postacute care settings are ongoing.^6,46 Recently, a national field test of observational data elements was conducted to assess pain among patients unable to communicate across postacute care settings (home health agencies, Inpatient Rehabilitation Facilities, Long-Term Care Hospitals, and Skilled Nursing Facilities) and it showed acceptable reliability.⁹ This underscores the need to implement standardized observational approaches to assess patients unable to self-report for pain in the HHC setting.^6,47,48 High-quality assessment of pain needs to incorporate multiple dimensions of pain assessment to inform better care planning to meet patient's needs.

The present study has several strengths worth highlighting. It provides a needed update on pain prevalence in HHC settings; it includes a comprehensive, recent national sample of HHC beneficiaries. This study improves upon prior knowledge because individual factors, including whether patients were able to self-report, were considered. Using national data our results confirm and expand upon the prior research findings. Our study has some limitations. First, although HHC clinicians are provided the same instructions with respect to how to collect OASIS data on pain level, variations in assessment and reporting might exist, and we lacked data on clinician type (RN or PT) conducting the assessment. Second, we lacked data on medications and thus could not evaluate pain management interventions received. Although we used OASIC-C, and the current OASIS is Version D, pain measures have remained unchanged.

In conclusion, older adults receiving HHC experience a high burden of severe pain. To our knowledge, this is the first study to examine the relationship between cognitive impairment and severe pain at a national level using OASIS data. Additional factors associated with severe pain in multivariable models suggest that vulnerable groups of older adults receiving HHC (those ≥85 years of age; Hispanic patients; patients with ADL dependency, diagnosis of ADRD, cognitive impairment, and those unable to self-report) were less likely to have severe pain documented. Because these findings likely reflect the underdiagnosis of pain in these groups, additional work to eliminate these disparities in pain assessment in HHC will require system-level interventions from administrators, researchers, and policymakers to ensure pain is being adequately identified and managed among those most at risk for inadequate treatment for severe pain.

Funding Information

Z.T.O was supported by a grant from the National Institute on Aging (5P30AG028741-07) awarded to the Claude D Pepper Older Americans Independence Center at the Icahn School of Medicine at Mount Sinai, and ZTO was also supported by the National Institute of Aging (NIA) of the National Institutes of Health (NIH) under Award Number U54AG063546 that funds NIA Imbedded Pragmatic Alzheimer's Disease and AD-Related Dementias Clinical Trials Collaboratory (NIA IMPACT Collaboratory).

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Centers for Medicare and Medicaid Services. Home Health Quality Measures. 2022. Available from: https://www.cms.gov/medicare/quality-initiatives-patient-assessment-instruments/homehealthqualityinits/home-health-quality-measures [Last accessed: December 9, 2022].

Medicare Payment Advisory Commission (MEDPAC). Report to Congress, Home Healthcare Services. March 2022.

Wang

, Monroe

, Simning

, et al. Pain management in home health care: Relationship with dementia and facility admissions. Pain Manag Nurs, 2021; 22(1):36–43.

Stubbs

, Binnekade

, Eggermont

, et al. Pain and the risk for falls in community-dwelling older adults: Systematic review and meta-analysis. Arch Phys Med Rehabil, 2014; 95(1):175–187.e9.

Covinsky

, Lindquist

, Dunlop

, et al. Pain, functional limitations, and aging. J Am Geriatr Soc, 2009; 57(9):1556–1561.

Dunbar

, Edelen

, McMullen

, et al. Development and testing of a standardized pain interview assessment for use in post-acute care. J Am Geriatr Soc, 2022; 70(4):1035–1046.

The AGS Panel on Persistent Pain in Older Persons. The management of persistent pain in older persons. J Am Geriatr Soc, 2002; 50(6 suppl):S205–S224; doi: 10.1046/j.1532-5415.50.6s.1.x.

Harris-Kojetin

, Sengupta

, Park-Lee

, et al. Long-term care providers and services users in the United States: Data from the National Study of Long-Term Care Providers, 2013–2014. Vital Health Stat, 2016;(38):x–xii; 1–105.

Shier

, Edelen

, McMullen

, et al. Standardized assessment of cognitive function, mood, and pain among patients who are unable to communicate. J Am Geriatr Soc, 2022; 70(4):1012–1022.

10.

Harkin

, Coates

, Brown

. Exploring ways to enhance pain management for older people with dementia in acute care settings using a Participatory Action Research approach. Int J Older People Nurs, 2022; 17(6):e12487.

11.

Cotton

, Lohman

, Brooks

, et al. Prevalence of and factors related to prescription opioids, benzodiazepines, and hypnotics among medicare home health recipients. Home Healthc Now, 2017; 35(6):304.

12.

Wang

, Cato

, Conwell

, et al. Pain treatment and functional improvement in home health care: Relationship with dementia. J Am Geriatr Soc, 2021; 69(12):3545–3556.

13.

Kang

, Reid

. Pain experiences in older adults with heart failure in the home healthcare setting after hospital discharge. Innov Aging, 2019; 3(Suppl 1):S262.

14.

Falvey

, Murphy

, Gill

, et al. Home health rehabilitation utilization among Medicare beneficiaries following critical illness. J Am Geriatr Soc, 2020; 68(7):1512–1519.

15.

Wang

, Cheng

, Kim

, et al. Pain and the Alzheimer's disease and related dementia spectrum in community-dwelling older Americans: A nationally representative study. J Pain Symptom Manage, 2022; 63(5):654–664.

16.

Centers for Medicare & Medicaid Services. Outcome and Assessment Information Set: OASIS-C Guidance Manual. Centers for Medicare & Medicaid Services: Baltimore, MD; 2009.

17.

Kinatukara

, Rosati

, Huang

. Assessment of OASIS reliability and validity using several methodological approaches. Home Healthc Serv Q, 2005; 24(3):23–38.

18.

O'Connor

, Davitt

. The Outcome and Assessment Information Set (OASIS): A review of validity and reliability. Home Healthc Serv Q, 2012; 31(4):267–301.

19.

Madigan

, Tullai-McGuinness

, Fortinsky

. Accuracy in the outcomes and assessment information set (OASIS): Results of a video simulation. Res Nurs Health, 2003; 26(4):273–283.

20.

Madigan

, Fortinsky

. Interrater reliability of the outcomes and assessment information set: Results from the field. Gerontologist, 2004; 44(5):689–692.

21.

Tullai-McGuinness

, Madigan

, Fortinsky

. Validity testing the outcomes and assessment information set (OASIS). Home Healthc Serv Q, 2009; 28(1):45–57.

22.

Burgdorf

, Amjad

. Cognitive impairment among medicare home health patients: Comparing available measures. Home Healthc Serv Q, 2022; 41(2):139–148.

23.

Forrester

, Mbrah

, Lapane

. A latent approach to understanding pain in nursing home residents who are unable to self-report pain. J Pain Res, 2021; 14:2283.

24.

Charlson

, Charlson

, Peterson

, et al. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. J Clin Epidemiol, 2008; 61(12):1234–1240.

25.

Services. Center for Medicare and Medicaid Services. Chronic Conditions Data Warehouse. Available from: http://www.ccwdata.org/chronic-conditions/index.htm [Last accessed: July 1, 2022].

26.

Lesaffre

, Marx

. Collinearity in generalized linear regression. Commun Stat Theory Methods, 1993; 22(7):1933–1952.

27.

Huang

C-CL

, Jou

Y-J

, Cho

H-J

. A new multicollinearity diagnostic for generalized linear models. J Appl Stat, 2016; 43(11):2029–2043.

28.

Fox

, Weisberg

An R companion to applied regression. Sage publications, 2018.

29.

Rene

, Baker

, Smith

, et al. The effect of race in older adults presenting for chronic pain management: A comparative study of black and white Americans. J Pain, 2003; 4(2):82–90.

30.

Morrison

, Jesdale

, Dube

, et al. Racial/ethnic differences in staff-assessed pain behaviors among newly admitted nursing home residents. J Pain Symptom Manage, 2021; 61(3):438–448.e3.

31.

Anderson

, Green

, Payne

. Racial and ethnic disparities in pain: Causes and consequences of unequal care. J Pain, 2009; 10(12):1187–1204.

32.

Mack

, Hunnicutt

, Jesdale

, et al. Non-Hispanic Black-White disparities in pain and pain management among newly admitted nursing home residents with cancer. J Pain Res, 2018; 11:753.

33.

Tsai

I-P

, Jeong

SY-S

, Hunter

. Pain assessment and management for older patients with dementia in hospitals: An integrative literature review. Pain Manag Nurs, 2018; 19(1):54–71.

34.

Boltz

, Resnick

, Kuzmik

, et al. Pain incidence, treatment, and associated symptoms in hospitalized persons with dementia. Pain Manag Nurs, 2021; 22(2):158–163.

35.

Hadjistavropoulos

, Herr

, Turk

, et al. An interdisciplinary expert consensus statement on assessment of pain in older persons. Clin J Pain, 2007; 23:S1–S43.

36.

Mandell

BF.

The fifth vital sign: A complex story of politics and patient care. Cleve Clin J Med, 2016; 83(6):400–401.

37.

Warden

, Hurley

, Volicer

. Development and psychometric evaluation of the Pain Assessment in Advanced Dementia (PAINAD) scale. J Am Med Dir Assoc, 2003; 4(1):9–15.

38.

Natavio

, McQuillen

, Dietrich

, et al. A comparison of the pain assessment checklist for seniors with limited ability to communicate (PACSLAC) and pain assessment in advanced dementia scale (PAINAD). Pain Manag Nurs, 2020; 21(6):502–509.

39.

Horgas

, Nichols

, Schapson

, et al. Assessing pain in persons with dementia: Relationships among the non-communicative patient's pain assessment instrument, self-report, and behavioral observations. Pain Manag Nurs, 2007; 8(2):77–85.

40.

Lautenbacher

, Sampson

, Pähl

, et al. Which facial descriptors do care home nurses use to infer whether a person with dementia is in pain?. Pain Med, 2017; 18(11):2105–2115.

41.

, Miller

, Lapane

, et al. The problem of assessment bias when measuring the hospice effect on nursing home residents' pain. J Pain Symptom Manage, 2003; 26(5):998–1009.

42.

Won

, Lapane

, Gambassi

, et al. Correlates and management of nonmalignant pain in the nursing home. J Am Geriatr Soc, 1999; 47(8):936–942.

43.

Tofthagen

, Halpenny

, Melendez

, et al. Evaluating the linguistic appropriateness and cultural sensitivity of a self-report system for spanish-speaking patients with cancer. Nurs Res Pract, 2014; 2014:702683.

44.

Nelson

Unequal treatment: Confronting racial and ethnic disparities in health care. J Natl Med Assoc, 2002; 94(8):666.

45.

Saliba

, Buchanan

. Making the investment count: Revision of the Minimum Data Set for nursing homes, MDS 3.0. J Am Med Dir Assoc, 2012; 13(7):602–610.

46.

Chen

, Edelen

, McMullen

, et al. Developing standardized patient assessment data elements for Medicare post-acute care assessments. J Am Geriatr Soc, 2022; 70(4):981–990.

47.

Langford

, Tauben

, Sturgeon

, et al. Treat the patient, not the pain: using a multidimensional assessment tool to facilitate patient-centered chronic pain care. J Gen Internal Med, 2018; 33(8):1235–1238.

48.

Gordon

DB.

Acute pain assessment tools: Let us move beyond simple pain ratings. Curr Opin Anesthesiol, 2015; 28(5):565–569.