Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Nov 17.
Published in final edited form as: Lancet Healthy Longev. 2023 Nov;4(11):e608–e617. doi: 10.1016/S2666-7568(23)00194-0

Association of neurocognitive disorders with morbidity and mortality in older adults undergoing major surgery in the USA: a retrospective, population-based, cohort study

Alexander T Abess 1, Stacie G Deiner 1, Alexandra Briggs 1, Elizabeth L Whitlock 1, Kristin E Charette 1, Vinca W Chow 1, Shahzad Shaefi 1, Pablo Martinez-Camblor 1, Alistair James O’Malley 1, Myles Dustin Boone 1
PMCID: PMC10654795  NIHMSID: NIHMS1942264  PMID: 37924842

Summary

Background

Neurocognitive disorders become increasingly common as patients age, and increasing numbers of surgical interventions are done on older patients. The aim of this study was to understand the clinical characteristics and outcomes of surgical patients with neurocognitive disorders in the USA in order to guide future targeted interventions for better care.

Methods

This retrospective cohort study used claims data for US Medicare beneficiaries aged 65 years and older with a record of inpatient admission for a major diagnostic or therapeutic surgical procedure between Jan 1, 2017, and Dec 31, 2018. Data were retrieved through a data use agreement between Dartmouth Hitchcock Medical Center and US Centers for Medicare and Medicaid Services via the Research Data Assistance Center. The exposure of interest was the presence of a pre-existing neurocognitive disorder as defined by diagnostic code within 3 years of index hospital admission. The primary outcome was mortality at 30 days, 90 days, and 365 days from date of surgery among all patients with available data.

Findings

Among 5 263 264 Medicare patients who underwent a major surgical procedure, 767 830 (14·59%) had a pre-existing neurocognitive disorder and 4 495 434 (85·41%) had no pre-existing neurocognitive disorder. Adjusting for demographic factors and comorbidities, patients with a neurocognitive disorder had higher 30-day (hazard ratio 1·24 [95% CI 1·23–1·25]; p<0·0001), 90-day (1·25 [1·24–1·26]; p<0·0001), and 365-day mortality (1·25 [1·25–1·26]; p<0·0001) compared with patients without a neurocognitive disorder.

Interpretation

Our findings suggest that the presence of a neurocognitive disorder is independently associated with an increased risk of mortality. Identification of a neurocognitive disorder before surgery can help clinicians to better disclose risks and plan for patient care after hospital discharge.

Funding

Department of Anesthesiology and Perioperative Medicine at Dartmouth Hitchcock Medical Center.

Introduction

Ageing populations are projected to strain health-care systems globally. Approximately one-third of all surgical procedures performed annually in the USA are in patients older than 65 years1 and, as the population ages, this proportion is expected to increase. Neurocognitive disorders become increasingly prevalent with age;2–4 consequently, a substantial number of surgical patients will present with coexisting neurocognitive disorders.5 Neurocognitive disorder is a broad term encompassing various diagnoses that manifest as cognitive function impairment and range from subtle, mild cognitive impairments to severe, disabling dementias.

Pre-existing neurocognitive disorder has been associated with higher postoperative complication rates, increased likelihood of discharge to facilities rather than home, and elevated mortality.6–8 However, the generalisability of previous studies is limited because of small study populations, short-term follow-up periods, or focus on specific surgery types or subcategories of neurocognitive disorder; a population-based, comprehensive view of the effect of neurocognitive disorder on a range of surgical outcomes is missing. To address this knowledge gap, we used the US Medicare Database, a federally funded US health insurance programme, to assess mortality and perioperative complications among adults aged 65 years and older undergoing major surgery. In 2021, approximately 60 million US residents (about 18% of the total US population at the time) were enrolled in Medicare. Medicare expenditure in the USA in 2021 exceeded US$900 billion, which is a substantial portion of the total $4 trillion spent on US health care altogether.9 In addition to mortality and complications, we aimed to describe the types of surgery and hospital and patient characteristics in this cohort. We hypothesised that the presence of a neurocognitive disorder would be associated with increased mortality and higher incidence of complications in older adults undergoing major surgery. Characterising the surgeries these patients undergo and their postoperative complications might support the design of future quality initiatives and research studies aimed at improving perioperative outcomes.

Methods

Study design and participants

This was a retrospective cohort study using claims data for US Medicare beneficiaries aged 65 years and older with an inpatient admission for a major diagnostic or therapeutic surgical procedure between Jan 1, 2017, and Dec 31, 2018. For patients with multiple admissions during this time, the first inpatient event was selected as the index admission. Procedural codes for major diagnostic or therapeutic procedures were determined using Healthcare Cost and Utilization Project methodology.10

All analyses adhered to a data use agreement with the Center for Medicare and Medicaid (CMS) through the Research Data Assistance Center at the University of Minnesota and the Committee for the Protection of Human Subjects at Dartmouth Hitchcock Medical Center. The Dartmouth Health Institutional Review Board approved this study, including waiver of participant consent in virtue of the nature of the study (secondary research on data previously collected; institutional review board identification: STUDY 02001078, notification letter dated June 16, 2021). The study was reported following the STROBE reporting guidelines.11 We accessed CMS data files dated Jan 1, 2014, to Dec 31, 2019, which included the 100% sample Medicare provider analysis and review file, outpatient, carrier, and master beneficiary summary files. All included patients were required to have available data for 3 years before admission (for assessment of the presence of neurocognitive disorder) and 1 year after admission (for observation of outcomes); for example, a patient with an index surgical hospital admission in 2017 required data from 2014 to 2018.

Procedures

Neurocognitive disorders are generally subdivided into major and minor classifications on the basis of their relative effect on an individual’s functional status. Major neurocognitive disorder results in impairment of independent living and function, whereas minor neurocognitive disorder does not generally affect independence. Dementia is the term most commonly associated with major neurocognitive disorder, whereas mild cognitive impairment is frequently interchangeable with minor neurocognitive disorder. Specific diagnostic criteria are enumerated elsewhere.12 Patients were classified as having a pre-existing neurocognitive disorder according to diagnostic codes including codes by the Diagnostic and Statistical Manual of Mental Illness, 5th edition, and code lists recommended by the Alzheimer’s Association Expert Task Force Consensus Statement and the European Delirium Association and American Delirium Society.13–16 These neurocognitive disorders included various dementias, cognitive impairments, delirium, and altered mental status. We chose to include altered mental status because it has been shown to commonly overlap with delirium in the older patient population.17,18 On the basis of these diagnoses, we used International Classification of Diseases (ICD) 9th revision and 10th revision codes to identify beneficiaries with a neurocognitive disorder.19,20 The transition in usage from ICD-9 to ICD-10 was mandated in the USA on Oct 1, 2015. The complete list of ICD codes used to assemble the neurocognitive disorder cohort is included in the appendix (p 1). To maximise sensitivity, we performed a 3-year look-back from the time of the index admission to identify neurocognitive disorder codes in line with previous studies.21–23

The hierarchical condition category, a risk-adjustment model used to estimate future health-care expenditures and provide a numeric summary of a patient’s comorbid conditions, was used to determine baseline health status. Other comorbidities (eg, BMI of 30 kg/m2 or above and cancer) and admission source (ie, home, emergency department, skilled nursing facility, or other institution) were also collected. Hospital characteristics (eg, type, size, and region) were recorded. Demographic factors such as age, sex, and race were collected. We chose the Social Deprivation Index (SDI) as a surrogate for socioeconomic status. The SDI is a geographical-based measure that incorporates data from seven categories of socioeconomic factors, such as proportion of people with income below the poverty level, proportion of people with less than 12 years of education, proportion of people living in high-density housing, proportion of people without a vehicle, and proportion of unemployed people. The measure has a direct, linear, positive relationship with social deprivation (in other words, as deprivation increases, so does the SDI score, which ranges from 1 to 100). SDI has been found to correlate with access to various health-care services and readmission rates in cognitive impairment research.24–27

Outcomes data were collected from the same CMS files, which include information such as date of death (if applicable), length of hospital stay, and discharge diagnoses. Mortality was obtained from the CMS master beneficiary summary file, length of stay and hospital complications were obtained from the CMS MedPAR file (inpatient data), and pre-existing neurocognitive disorder diagnoses were identified using CMS outpatient and carrier files.

Outcomes

The primary outcome was overall survival at 30 days, 90 days, and 365 days from index surgery. Secondary outcomes included discharge destination (eg, home, rehabilitation, or skilled nursing facilities), readmission rates, complications (eg, delirium, cardiovascular events [myocardial infarction, pulmonary embolism, stroke, and deep vein thrombosis], infection [pneumonia, urinary tract infection, superficial site infection, wound infection, and sepsis], reoperation, and renal insufficiency. We also described as a prespecified secondary outcome the types of surgery that patients with neurocognitive disorder underwent compared with patients without neurocognitive disorder as well as hospital type and region between the two groups.

Statistical analysis

Descriptive statistics of the data were collected. The distributions of categorical data were presented as frequencies and proportions. The strength of the association between neurocognitive disorder and the primary and secondary outcomes was measured as hazard ratios (HRs) when the outcome was time-dependent (ie, time-to-death and time-to-readmission), odds ratios (ORs) when the outcome was binary, and as regression coefficients interpreted as differences in expected mean outcomes for continuous variables. The probability of death within the first year was estimated using the Kaplan-Meier estimator. For statistical inferential summaries, we provide the unadjusted and adjusted coefficients with 95% CIs. Different models are provided to represent how the covariates modify the association of neurocognitive disorder with the outcome of interest. We chose three different models to explore the effect of different covariates on the level of association between neurocognitive disorder and mortality: an unadjusted model, model one (adjusting for age, sex, and race), and model two (adjusting for age, sex, race, hierarchical condition category, and SDI). Secondary outcomes were evaluated with a model that was adjusted for demographic variables (ie, age, sex, race, and SDI) and patient variables (ie, BMI of 30 kg/m2 or above, history of cancer, hierarchical condition category, admission source from nursing facility, and emergency admission type). Multiple-comparison Bonferroni correction was applied on the CIs for the secondary outcomes. All analyses were performed using R statistical software, version 4.0.1, with two-sided p<0·05 considered statistically significant.

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

Between Jan 1, 2017, and Dec 31, 2018, 5 263 264 eligible patients were admitted to hospital as inpatients for a major diagnostic or therapeutic surgical procedure (figure 1). Of these patients, 767 830 (14·59%) had at least one neurocognitive disorder diagnosis: 413 413 (53·84%) had a single diagnosis, 154 485 (20·12%) had two diagnoses (eg, Alzheimer’s disease with late onset in addition to dementia without behavioural disturbance), 79 365 (10·34%) had three diagnoses, and 120 567 (15·70%) had four or more diagnoses. Thus, the total number of neurocognitive disorder diagnoses exceeded the number of patients with a neurocognitive disorder diagnosis (appendix p 2). Baseline characteristics of the cohort are shown in table 1. Patients with a neurocognitive disorder diagnosis tended to be older and were more likely to be female than those without a neurocognitive disorder. Median hierarchical condition category risk score and SDI were higher in patients with a neurocognitive disorder than in those without a neurocognitive disorder. A greater proportion of patients with neurocognitive disorder were admitted from a skilled nursing facility or through the emergency department, compared with patients without a neurocognitive disorder.

Figure 1:

Figure 1:

Open in a new tab

Inclusion flow diagram

Table 1:

Characteristics of patients with and without a neurocognitive disorder

Total participants (n=5 263 264) Neurocognitive disorder absent (n=4 495 434) Neurocognitive disorder present (n=767 830)
Age, years 74 (69–80) 74 (69–79) 79 (72–86)
Sex
 Female 2 817 877 (53·54%) 2 375 264 (52·83%) 442 613 (57·64%)
 Male 2 445 387 (46·46%) 2 120 170 (47·17%) 325 217 (42·36%)
BMI ≥30 kg/m2 711 623 (13·52%) 635 205 (14·13%) 76 418 (9·95%)
History of cancer 753 226 (14·31%) 608 374 (13·53%) 144 852 (18·87%)
Hierarchical condition category score 0·97 (0·33–2·26) 0·85 (0·29–1·99) 1·99 (0·97–3·58)
Social Deprivation Index score 40 (19–63) 39 (19–63) 42 (21–66)
Race
 Unknown 79 180 (1·50%) 73 224 (1·63%) 5956 (0·78%)
 White 4 520 598 (85·89%) 3 865 690 (85·99%) 654 908 (85·29%)
 Black 405 459 (7·70%) 335 815 (7·47%) 69 644 (9·07%)
 Other 77 482 (1·47%) 68 380 (1·52%) 9102 (1·19%)
 Asian 67 261 (1·28%) 57 138 (1·27%) 10 123 (1·32%)
 Hispanic 90 062 (1·71%) 76 102 (1·69%) 13 960 (1·82%)
 North American Native 23 222 (0·44%) 19 085 (0·42%) 4137 (0·54%)
Admission from skilled nursing facility 47 268 (0·90%) 18 475 (0·41%) 28 793 (3·75%)
Emergency admission 1705 652 (32·41%) 1 331 196 (29·61%) 374 456 (48·77%)
Open in a new tab

Data are median (IQR) or n (%).

The observed mortality difference between patients with a neurocognitive disorder and patients without a neurocognitive disorder is in figure 2A. Before adjustment, the presence of a neurocognitive disorder compared with no neurocognitive disorder was associated with increased mortality at 30 days (7·16% [95% CI 7·10–7·22] vs 3·06% [3·05–3·08], HR 2·38 [95% CI 2·36–2·40]; p<0·0001), 90 days (13·92% [13·85–14·00] vs 5·57% [5·55–5·59], 2·60 (2·58–2·62); p<0·0001), and 1 year (26·41% [26·32–26·51] vs 10·66% [10·63–10·69], 2·71 [2·70–2·72]; p<0·0001). When adjusting for age, sex, and race (model one), neurocognitive disorder was associated with increased mortality at 30 days (HR 1·86 [95% CI 1·84–1·88]; p<0·0001), 90 days (2·04 [2·01–2·05]; p<0·0001), and 365 days (2·16 [2·15–2·17]; p<0·0001; figure 2B). When adjusting for age, sex, race, hierarchical condition category, and SDI (model two), neurocognitive disorder remained associated with increased mortality at 30 days (HR 1·24 [95% CI 1·23–1·25]; p<0·0001), 90 days (1·25 [1·24–1·26]; p<0·0001), and 1 year (1·25 [1·25–1·26]; p<0·0001; figure 2B).

Figure 2: Mortality risk over time and HRs at 30 days, 90 days, and 365 days for unadjusted and adjusted models.

Figure 2:

Open in a new tab

(A) Kaplan-Meier curve of observed mortality during follow-up in patients with a neurocognitive disorder and patients without a neurocognitive disorder. (B) Forest plots of HRs in the unadjusted and adjusted models for mortality at 30 days, 90 days, and 365 days of follow-up in patients with and without a neurocognitive disorder. Model one is adjusted for age, sex, and race. Model two is adjusted for age, sex, race, hierarchical condition category, and Social Deprivation Index. HR=hazard ratio.

Adjusted analyses of postoperative complications, readmission rates, and discharge destinations for patients with and without a neurocognitive disorder are shown in table 2. The models were adjusted by age, sex, race, BMI of 30 kg/m2 or above, cancer history, hierarchical condition category score, SDI, admission from a nursing facility, and emergency admission. ORs for a diagnosis of delirium, stroke, wound infection, and urinary tract infection during index admission were higher in patients with a neurocognitive disorder than in patients without a neurocognitive disorder (table 2).

Table 2:

Postoperative outcomes in patients with and without a neurocognitive disorder

Neurocognitive disorder absent (n=4 495 434) Neurocognitive disorder present (n=767 830) Adjusted effect size (95% CI) p value
Post-surgical admission to intensive care unit 1 239 383 (27·57%) 254 281 (33·11%) 0·86 (0·85 to 0·87) <0·0001
Length of hospital stay, days 3 (2–6) 4 (3–8) –0·38 (–0·40 to –0·36) <0·0001
Discharge destination
 Home 2 145 918 (47·72%) 197 293 (25·69%) 0·64 (0·63 to 0·65) <0·0001
 Skilled nursing facility 840 247 (18·69%) 306 379 (39·90%) 1·73 (1·71 to 1·74) <0·0001
 Rehabilitation facility 216 161 (4·81%) 61 551 (8·02%) 1·22 (1·20 to 1·24) <0·0001
Readmission within 30 days 515 999 (11·48%) 157 998 (20·58%) 1·07 (1·06 to 1·08) <0·0001
Readmission within 90 days 847 927 (18·86%) 252 956 (32·94%) 1·08 (1·07 to 1·09) <0·0001
Complications during inpatient stay
 Delirium 41 459 (0·92%) 22 075 (2·87%) 1·96 (1·91 to 2·01) <0·0001
 Pneumonia 110 178 (2·45%) 34 207 (4·46%) 0·93 (0·91 to 0·95) <0·0001
 Pulmonary embolism 23 231 (0·52%) 4445 (0·58%) 0·63 (0·60 to 0·66) <0·0001
 Deep vein thrombosis 53 004 (1·18%) 13 058 (1·70%) 0·79 (0·76 to 0·80) <0·0001
 Stroke 283 789 (6·31%) 80 395 (10·47%) 1·38 (1·36 to 1·40) <0·0001
 Cardiac arrest 26 278 (0·58%) 5993 (0·78%) 0·76 (0·73 to 0·80) <0·0001
 Myocardial infarction 306 990 (6·83%) 44 728 (5·83%) 0·66 (0·64 to 0·66) <0·0001
 Renal insufficiency 1 133 055 (25·20%) 301 110 (39·21%) 1·08 (1·07 to 1·09) <0·0001
 Wound infection 2940 (0·07%) 1190 (0·15%) 1·28 (1·14 to 1·42) <0·0001
 Superficial site infection 21 026 (0·47%) 4255 (0·55%) 0·79 (0·74 to 0·83) <0·0001
 Urinary tract infection 166 145 (3·70%) 69 819 (9·09%) 1·43 (1·40 to 1·45) <0·0001
 Sepsis 118 467 (2·64%) 40 247 (5·24%) 0·93 (0·91 to 1·95) <0·0001
 Need for reoperation 621 (0·01%) 143 (0·02%) 1·19 (0·88 to 1·60) 0·93
Open in a new tab

Data are n (%) or median (IQR), except where otherwise stated. Effect sizes are adjusted by all covariates reported in table 1. The adjusted effect size for length of stay is reported as the regression coefficient. For readmission rates (within 30 days and 90 days), the effect size is reported as a hazard ratio; all other effect sizes are odds ratios.

The five most common surgical procedures for patients with neurocognitive disorder were treatment of a hip or femur fracture (107 647 [14·02%] of 767 830), followed by hip arthroplasty (85 446 [11·13%]), knee arthroplasty (46 039 [6·00%]), percutaneous transluminal coronary angioplasty (41 691 [5·43%]), and vascular surgery (33 040 [4·30%]; figure 3). By comparison, the five most common procedures for patients without neurocognitive disorder were knee arthroplasty (671 207 [14·93%] of 4 495 434), followed by hip arthroplasty (503 079 [11·19%]), percutaneous transluminal coronary angioplasty (297 936 [6·63%]), spinal fusion (247 785 [5·51%]), and treatment of hip or femur fracture (226 755 [5·04%]). Rates for other procedure types were similar across the two groups.

Figure 3: Most frequent surgery types for patients with and without a neurocognitive disorder.

Figure 3:

Open in a new tab

The figure shows all surgery types that accounted for at least 1% of surgeries. Surgery type is based on the Clinical Classification Software scheme from the Agency for Healthcare Research and Quality.

We compared hospital and regional characteristics for patients with and without neurocognitive disorder (appendix p 3). The hospital size and owner type were similar across the two groups. Patients with a neurocognitive disorder were slightly more common than patients without a neurocognitive disorder in rural and southern hospitals. Hospitals outside of the continental USA had fewer patients with a neurocognitive disorder than hospitals in the continental USA (appendix p 3).

Discussion

In this retrospective cohort study of Medicare beneficiaries aged 65 years or older undergoing major surgical procedures, we observed a significant association between the presence of a neurocognitive disorder and mortality at 30 days, 90 days, and 365 days. These results remained consistent after adjustments for age, comorbidities, and socioeconomic status. Our findings offer population-based estimates of mortality rates in older adults with neurocognitive disorder after undergoing major surgery.

A recent study involving more than 5000 community-living older adults in the USA also showed an increased risk of death after major surgery in patients with dementia and frailty.28 Our study substantially extends these findings by showing a similar risk of mortality after major surgery in patients with all types of neurocognitive disorder and across a much larger segment of the US population. In addition, we observed that, compared with patients without a neurocognitive disorder, patients with a neurocognitive disorder had higher rates of perioperative complications such as delirium, stroke, infection, reoperation, and readmission, and were less likely to be discharged to their homes. These different outcomes are likely to be meaningful for care providers, caregivers, and patients, who often prioritise quality of life and functional independence over quantity of life.29

An unexpected finding was the lower adjusted risk for some complications (eg, cardiac arrest, myocardial infarction, pulmonary embolism, and superficial site infections) in patients with neurocognitive disorder compared with patients without neurocognitive disorder, which we hypothesise might reflect an epiphenomenon in which symptomatic reporting and clinical detection of these complications in perioperative patients with neurocognitive disorder is hindered. Whether this lower rate of observed complications could potentially be a source of increased mortality because of unrealised treatment for unrecognised conditions is unknown, and further investigation in this topic seems warranted.

In this study, 14·59% of Medicare beneficiaries who had a major surgical procedure had one or more neurocognitive disorder diagnoses. These rates of neurocognitive disorder almost certainly represent an underdiagnosis of these conditions. Only 156 154 (2·97%) of the more than 5·2 million patients included in this study had a diagnosis of cognitive impairment. Previous estimates have indicated that the prevalence of cognitive impairment is several times higher in surgical patients older than 65 years than what we observed in this study.6,30 CMS has implemented a recommendation that older patients be assessed for cognitive impairment as part of an annual wellness visit.31 Regrettably, this potential benefit remains underused: only 20% of Medicare fee-for-service beneficiaries received a screening examination in 2016, suggesting that older adults who are eligible for cognitive testing via this mechanism continue to be missed.32 Beyond issues of accessibility, the effectiveness of the Medicare annual wellness visit in detecting cognitive impairment has been scrutinised.33 CMS recommends that providers detect cognitive impairment during the annual wellness visit on the basis of direct observation or any concerns expressed by patients or their family or friends. However, CMS does not endorse the use of a specific screening tool.31 Furthermore, in 2020 the US Preventive Services Task Force reiterated its previous finding that the evidence to justify routine cognitive screening in older patients is insufficient.34 Their conclusion that there is insufficient evidence to justify routine cognitive screening remains a subject of debate, as the American Academy of Neurology and the Alzheimer’s Association both advocate for the use of a screening tool to assess for cognitive impairment.35,36 Uncertainty regarding the relative benefits and harms of cognitive screening arises from the insufficient data describing impact on outcomes that are meaningful to patients, caregivers, and society.37

This study suggests the need for preoperative identification of cognitive impairment to inform patient and provider discussions surrounding treatment and recovery. As the global population ages, more adults with neurocognitive disorder will undergo major surgery, highlighting the need to focus efforts on early identification of neurocognitive disorders and perioperative risk mitigation.38 The American Society of Anesthesiologists, the European Society of Anaesthesiology, the Royal College of Anaesthetists, the American College of Surgeons, and the American Geriatrics Society all have recommended preoperative cognitive screening for patients older than 65 years before major surgery, but the implementation of these recommendations appears to be limited. This study contributes substantially to the supporting evidence for these recommendations and suggests that collecting intermediate and longer term metrics in patients with neurocognitive impairment will be important to improve care.

Multidisciplinary preoperative programmes aimed at risk stratification of older adults at high risk for perioperative complications have shown success in enhancing postoperative and quality-of-life outcomes.39,40 The American College of Surgeons Geriatric Surgery Verification programme, which aims to optimise the surgical care of older adults through implementation of standards spanning both the preoperative and postoperative periods, includes preoperative cognitive assessment as one element of a comprehensive screen for vulnerability.41 Implementation of this programme has been shown to reduce the length of inpatient stay.42

Delirium prevention strategies have been shown in international settings to effectively reduce the incidence and severity of delirium after surgery.43 Our study showed that patients with a neurocognitive disorder had a higher risk of delirium after surgery compared with patients without neurocognitive disorder (OR 1·96 [95% CI 1·91–2·01]), which indicates that this patient population is an important group on which to focus these preventive efforts.

Our study also identified the most common surgical procedures done in older adults in the USA, which might inform targeted subgroup interventions to improve outcomes. For example, we found that fractures are a primary driver of hospitalisations associated with a major procedure in patients with a neurocognitive disorder. Public health prevention programmes designed to reduce fracture risk in older adults might substantially reduce hospitalisation rates.

Our study has several limitations. First, the inherent limitations of administrative claims data must be considered. These data are primarily derived to serve billing and administrative purposes and not clinical research, which might result in inaccuracies or underreporting of specific diagnoses and procedures. Administrative staff examine clinical documentation of procedures and diagnoses performed by clinicians that have varying levels of training (although the billing provider is ultimately responsible for accuracy), and this clinical documentation is then converted to administrative codes. Additionally, the potential exists for a lack of coding granularity, miscoding, or undercoding due to the vast number of codes available, leading to the possibility of misclassification bias or affecting the depth of our analysis. Overarchingly, the retrospective nature of these data with a varied degree of confounding variable data reporting limits the ability to infer causation.

Second, misclassification of the exposure might have occurred. Some patients with a neurocognitive disorder are likely not to have an ICD-coded diagnosis—for example, previous research has suggested that approximately half of all patients with dementia do not have a diagnosis,44–49 and using Medicare claims data to identify dementia has been shown to miss about half of clinically diagnosed patients.22 As such, our results are likely to be imprecise due to misclassification of patients with a mild neurocognitive disorder in the subgroup of patients deemed not to have a neurocognitive disorder. The degree to which this possible misclassification alters our results is unknown, and potential differences in outcomes might be greater as well as lesser. The presence of falsely classified patients (those with actual disease that evades detection) in the cohort without neurocognitive disorder could cause a relative worsening of outcomes in this cohort (thus minimising the difference between the two groups). The same misclassification could also possibly cause the outcomes to appear worse than reality in the group with a neurocognitive disorder than in the group without a neurocognitive disorder because only the most severe cases of neurocognitive disorder (with potentially the worst outcomes) are currently counted in that group. However, without knowing the actual crossover and effect of the varying degrees of misclassified patients, knowing in which direction or to what degree our results might be skewed is impossible. One of our strategies to increase sensitivity to patients with a presumably undiagnosed mild neurocognitive disorder was to include non-specific diagnoses, such as altered mental status and unspecified signs and symptoms of cognitive function and awareness as neurocognitive disorders. As such, it should be noted that the neurocognitive diagnoses chosen for inclusion are based on ICD codes and are distinct from those described in the perioperative neurocognitive disorder literature.5

Third, our ability to examine subtypes of neurocognitive disorder in relation to outcome is limited for several reasons: the transition from ICD 9th revision to ICD 10th revision; the evolving terminology and diagnostic patterns of neurocognitive disorders; and constrained diagnostic nuance, accuracy, and potential duplication from using administrative claims data. For example, we chose not to use the number of neurocognitive disorder diagnoses because the same clinical entity could conceivably be diagnosed with multiple codes (for example, either due to different coding practices or as the disease progresses and changes in presentation). Similarly, we chose not to attempt discrimination between minor and major neurocognitive disorders because the terminology, diagnoses, and coding of neurocognitive disorders has changed substantially over the past 10 years. For these reasons, we deemed it most appropriate to use a dichotomous exposure variable (presence or absence of neurocognitive disorder), which offered a straightforward approach to assess the effect of neurocognitive disorder on our outcomes of interest without the complications that come with grading or stratifying on the basis of the number or severity of diagnoses. This method simplified the interpretation of results, making this interpretation more accessible and providing clear and actionable insights.

Fourth, the composition of our cohort might limit the generalisability of our findings. Medicare claims data are limited to the US population and under-represent patients younger than 65 years, those covered by Medicare Advantage, and those receiving dual Medicare and Medicaid (a state-run insurance plan for younger people, often with low-income, or those with qualifying medical conditions). About 12% of people aged 65 years and older in the USA have Medicare as their sole insurance provider.9 Another 40% of US adults older than 65 years also have private insurance, which might be combined with Medicare.9 Medicare Advantage programmes are offered by Medicare-approved companies that must follow rules set by Medicare and cover another 40% of US people older than 65 years. However, of note, demographic and socioeconomic characteristics of patients receiving Medicare and Medicare Advantage are similar.50

We chose to exclude data from 2020 to 2022 because of disruptions in health-care delivery related to the COVID-19 pandemic. Additionally, we cannot exclude the possibility that unmeasured or incorrectly coded confounding variables affected our modelling. For example, we did not include surgery type as a covariate in our models. Lastly, our 1-year outcome data are limited to mortality.

Future research should be directed towards other patient-centred outcomes such as functional independence and quality of life, which could enhance preoperative decision making. Focused studies for specific surgery types would also be of value, and orthopaedic surgery examination in particular might have a high clinical impact.

Strategies to incorporate neurocognitive disorder screening should be developed so that appropriate risk stratification and supported decision making can occur before surgical interventions. Ideally, neurocognitive disorder screening can occur even in the setting of urgent and emergent procedures to inform preoperative discussions of risk and benefit. Counselling should include the likelihood of skilled nursing admission (at least temporarily) and 1-year mortality, which in many cases providers can tailor to their patient and procedure using tools such as the American College of Surgeons risk calculator.51 Identifying the presence of preoperative neurocognitive disorder can inform more accurate supported decision making. Further incorporating the use of patient decision aids has been shown to be associated with several benefits, including more accurate risk perceptions, increased alignment of decisions with patients’ values, reduced internal conflict for patients and families, and decreased pursuit of major elective surgery.52

Supplementary Material

1

Research in context.

Evidence before this study

Estimates suggest that, by 2030, a sixth of the world’s population will be older than 60 years. One in every three surgeries in the USA is done in patients older than 65 years. Neurocognitive disorders tend to become increasingly common as people age; consequently, increasing numbers of patients requiring surgery are likely to have concurrent neurocognitive conditions. We searched PubMed from Jan 1, 2012 to Dec 31, 2022 for publications in English with the terms “neurocognitive disorder”, “surgery”, and “outcomes” to assess the impact of pre-existing neurocognitive disorders on surgical outcomes. Results were restricted to studies on humans aged 65 years and older, and review articles and meta-analyses were excluded. We found 283 studies, most of which were limited by small numbers of participants and specific surgery types. Surgical patients with a neurocognitive disorder in selected populations have been previously shown to have worse outcomes.

Added value of this study

This study uses claims data from the US Medicare programme to provide broad, population-based evidence that older patients with pre-existing neurocognitive disorders have a greater risk of mortality and are less likely to be discharged home following major surgery than patients without neurocognitive disorders. This association persists even after adjusting for age, comorbidities, and socioeconomic status. Importantly, this study also provides a description of the common types of major surgery that older patients with neurocognitive disorders undergo, thus highlighting high-yield areas for care improvement efforts.

Implications of all the available evidence

The presence of a neurocognitive disorder is associated with markedly worse outcomes in older surgical patients. As such, identification of a neurocognitive disorder before surgery is important to allow accurate risk disclosure and care planning at the individual patient level. Ageing populations are projected to strain health-care systems globally. Active screening for neurocognitive disorders might offer an opportunity for health systems to identify particularly vulnerable patients and support efficient deployment of health-care resources.

Acknowledgments

The work described herein was funded by the Department of Anesthesiology and Perioperative Medicine at Dartmouth Hitchcock Medical Center. SS’s time for this work was supported by the following National Institute of Health grants: R03AG060179, K08 GM134220, R01DK126685, and R01DK125786. ELW receives funding from the US National Institutes of Health’s National Institute on Aging grant R01AG079263.

AB reports compensation from the American College of Surgeons for her previous work related to the Geriatric Surgical Verification programme; funding from the Susan and Levy Health Care Delivery Incubator at Dartmouth for implementation of a geriatric surgery programme; and speaker’s fees for lectures related to geriatric surgery. ELW reports research funding from the US National Institutes of Health (National Institute on Aging). SS reports research funding from the US National Institutes of Health (National Institute on Aging and National Institute of General Medical Sciences). SGD reports payment for expert testimony; is a Director of the American Board of Anesthesiology; and is Chair of the American Society of Anesthesiology Brain Health Subcommittee. ATA is a member of the Multicenter Perioperative Outcomes Group brain health working group; is co-inventor on several medical device patent applications (none related to brain health or neurocognitive disorders); has previously started a medical device company directed towards developing sensor-enhanced needles; is co-founder of a non-profit medical innovation group at Dartmouth Health; and is an adviser to Simbex Corporation, a medical technology development company, but receives no remuneration for this work and has not consulted on any technologies related to the subject matter of this manuscript.

Footnotes

Data sharing

The US Centers for Medicare and Medicaid Services (CMS) have strict data security requirements in place that prohibit data sharing without approval. A signed data use agreement with CMS is required before data can be accessed. The investigators are willing to support reasonable inquiries and applications for data access, but final approval must be granted by CMS. Requests for access to CMS data can be made through https://www.resdac.org. The corresponding author (alexander.t.abess@hitchcock.org) should be contacted for further information on how to request access to the data.

Declaration of interests

All other authors declare no competing interests.

References

  • 1.Hall MJ, Schwartzman A, Zhang J, Liu X. Ambulatory surgery data from hospitals and ambulatory surgery centers: United States, 2010. https://www.cdc.gov/nchs/data/nhsr/nhsr102.pdf (accessed Oct 10, 2023). [PubMed]
  • 2.Holsinger T, Deveau J, Boustani M, Williams JW Jr. Does this patient have dementia? JAMA 2007; 297: 2391–404. [DOI] [PubMed] [Google Scholar]
  • 3.Roberts RO, Geda YE, Knopman DS, et al. The incidence of MCI differs by subtype and is higher in men: the Mayo Clinic Study of Aging. Neurology 2012; 78: 342–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Manly JJ, Tang MX, Schupf N, Stern Y, Vonsattel JPG, Mayeux R. Frequency and course of mild cognitive impairment in a multiethnic community. Ann Neurol 2008; 63: 494–506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Evered L, Silbert B, Knopman DS, et al. Recommendations for the nomenclature of cognitive change associated with anaesthesia and surgery—2018. Br J Anaesth 2018; 121: 1005–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Robinson TN, Wu DS, Pointer LF, Dunn CL, Moss M. Preoperative cognitive dysfunction is related to adverse postoperative outcomes in the elderly. J Am Coll Surg 2012; 215: 12–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hu CJ, Liao CC, Chang CC, Wu CH, Chen TL. Postoperative adverse outcomes in surgical patients with dementia: a retrospective cohort study. World J Surg 2012; 36: 2051–58. [DOI] [PubMed] [Google Scholar]
  • 8.Seitz DP, Gill SS, Gruneir A, et al. Effects of dementia on postoperative outcomes of older adults with hip fractures: a population-based study. J Am Med Dir Assoc 2014; 15: 334–41. [DOI] [PubMed] [Google Scholar]
  • 9.Cha AE, Cohen RA. Demographic variation in health insurance coverage: United States, 2020. https://www.cdc.gov/nchs/data/nhsr/nhsr169.pd (accessed Oct 9, 2023). [PubMed]
  • 10.Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project user support. https://hcup-us.ahrq.gov/toolssoftware/procedureicd10/procedure_icd10.jsp (accessed March 21, 2022).
  • 11.Gharaibeh A, Koppikar S, Bonilla-Escobar J. F. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) in the International Journal of Medical Students. Int J Med Stud 2014; 2: 36–37. [Google Scholar]
  • 12.American Psychiatric Association. Diagnostic and statistical manual of mental disorders: DSM-5, 5th edn. Washington, DC: American Psychiatric Association, 2022. [Google Scholar]
  • 13.First M. New ICD-10-CM codes for neurocognitive disorders effective October 1. American Psychiatric Association. 10.1176/appi.pn.2022.10.10.34 (accessed Oct 9, 2023). [DOI] [Google Scholar]
  • 14.Johnson LM. ICD10 monitor: Alzheimer’s, up close and personal. https://www.icd10monitor.com/alzheimer-s-up-close-and-personal (accessed Nov 22, 2020).
  • 15.Sachdev PS, Blacker D, Blazer DG, et al. Classifying neurocognitive disorders: the DSM-5 approach. Nat Rev Neurol 2014; 10: 634–42. [DOI] [PubMed] [Google Scholar]
  • 16.Alzheimer’s Association Expert Task Force. Consensus statement on CMS proposed billing code for the assessment and care planning for individuals with cognitive impairment. 2016. http://act.alz.org/site/DocServer/Taskforce_Consensus_Statement_FINAL.pdf?docID=51841 (accessed Oct 9, 2023).
  • 17.European Delirium Association. The DSM-5 criteria, level of arousal and delirium diagnosis: inclusiveness is safer. BMC Med 2014; 12: 141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Han JH, Zimmerman EE, Cutler N, et al. Delirium in older emergency department patients: recognition, risk factors, and psychomotor subtypes. Acad Emerg Med 2009; 16: 193–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.WHO. International classification of diseases: [9th] ninth revision, basic tabulation list with alphabetic index. Geneva: World Health Organization, 1978. [Google Scholar]
  • 20.WHO. ICD-10: international statistical classification of diseases and related health problems: tenth revision. 2nd edn, Geneva: World Health Organization, 2004. [Google Scholar]
  • 21.Lee E, Gatz M, Tseng C, et al. Evaluation of Medicare claims data as a tool to identify dementia. J Alzheimers Dis 2019; 67: 769–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Zhu CW, Ornstein KA, Cosentino S, Gu Y, Andrews H, Stern Y. Misidentification of dementia in Medicare claims and related costs. J Am Geriatr Soc 2019; 67: 269–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Taylor DH Jr, Fillenbaum GG, Ezell ME. The accuracy of Medicare claims data in identifying Alzheimer’s disease. J Clin Epidemiol 2002; 55: 929–37. [DOI] [PubMed] [Google Scholar]
  • 24.Butler DC, Petterson S, Phillips RL, Bazemore AW. Measures of social deprivation that predict health care access and need within a rational area of primary care service delivery. Health Serv Res 2013; 48: 539–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Phillips RL, Liaw W, Crampton P, et al. How other countries use deprivation indices—and why the United States desperately needs one. Health Aff (Millwood) 2016; 35: 1991–98. [DOI] [PubMed] [Google Scholar]
  • 26.Liaw W, Krist AH, Tong ST, et al. Living in “cold spot” communities is associated with poor health and health quality. J Am Board Fam Med 2018; 31: 342–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Robert Graham Center. Social Deprivation Index (SDI). https://www.graham-center.org/maps-data-tools/social-deprivation-index.html (accessed Dec 18, 2022).
  • 28.Gill TM, Vander Wyk B, Leo-Summers L, Murphy TE, Becher RD. Population-based estimates of 1-year mortality after major surgery among community-living older US adults. JAMA Surg 2022; 157: e225155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Higginson IJ, Daveson BA, Morrison RS, et al. Social and clinical determinants of preferences and their achievement at the end of life: prospective cohort study of older adults receiving palliative care in three countries. BMC Geriatr 2017; 17: 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Culley DJ, Flaherty D, Fahey MC, et al. Poor performance on a preoperative cognitive screening test predicts postoperative complications in older orthopedic surgical patients. Anesthesiology 2017; 127: 765–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.CMS Annual Wellness Visit. https://www.ecfr.gov/current/title-42/chapter-IV/subchapter-B/part-410/subpart-B/section-410.15 (accessed Oct 9, 2023).
  • 32.Ganguli I, Souza J, McWilliams JM, Mehrotra A. Trends in use of the US Medicare annual wellness visit, 2011–2014. JAMA 2017; 317: 2233–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Fowler NR, Campbell NL, Pohl GM, et al. One-year effect of the Medicare annual wellness visit on detection of cognitive impairment: a cohort study. J Am Geriatr Soc 2018; 66: 969–75. [DOI] [PubMed] [Google Scholar]
  • 34.Patnode CD, Perdue LA, Rossom RC, et al. Screening for cognitive impairment in older adults: updated evidence report and systematic review for the US Preventive Services Task Force. JAMA 2020; 323: 764–85. [DOI] [PubMed] [Google Scholar]
  • 35.Petersen RC, Lopez O, Armstrong MJ, et al. Practice guideline update summary: mild cognitive impairment: report of the guideline development, dissemination, and implementation subcommittee of the American Academy of Neurology. Neurology 2018; 90: 126–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Cordell CB, Borson S, Boustani M, et al. Alzheimer’s Association recommendations for operationalizing the detection of cognitive impairment during the Medicare annual wellness visit in a primary care setting. Alzheimers Dement 2013; 9: 141–50. [DOI] [PubMed] [Google Scholar]
  • 37.Owens DK, Davidson KW, Krist AH, et al. Screening for cognitive impairment in older adults: US Preventive Services Task Force recommendation statement. JAMA 2020; 323: 757–63. [DOI] [PubMed] [Google Scholar]
  • 38.Becher RD, Vander Wyk B, Leo-Summers L, Desai MM, Gill TM. The incidence and cumulative risk of major surgery in older persons in the United States. Ann Surg 2023; 277: 87–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Zietlow KE, Wong S, Heflin MT, et al. Geriatric preoperative optimization: a review. Am J Med 2022; 135: 39–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.McDonald SR, Heflin MT, Whitson HE, et al. Association of integrated care coordination with postsurgical outcomes in high-risk older adults: the Perioperative Optimization of Senior Health (POSH) initiative. JAMA Surg 2018; 153: 454–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.American College of Surgeons. Optimal resources for geriatric surgery: 2019. Standards. https://www.facs.org/media/f10eka54/geriatricsv_standards.pdf (accessed March 30, 2023).
  • 42.Jones TS, Jones EL, Richardson V, et al. Preliminary data demonstrate the Geriatric Surgery Verification program reduces postoperative length of stay. J Am Geriatr Soc 2021; 69: 1993–99. [DOI] [PubMed] [Google Scholar]
  • 43.Wang YY, Yue JR, Xie DM, et al. Effect of the tailored, family-involved hospital elder life program on postoperative delirium and function in older adults: a randomized clinical trial. JAMA Intern Med 2020; 180: 17–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Boustani M, Peterson B, Hanson L, Harris R, Lohr KN. Screening for dementia in primary care: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2003; 138: 927–37. [DOI] [PubMed] [Google Scholar]
  • 45.Bradford A, Kunik ME, Schulz P, Williams SP, Singh H. Missed and delayed diagnosis of dementia in primary care: prevalence and contributing factors. Alzheimer Dis Assoc Disord 2009; 23: 306–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Kotagal V, Langa KM, Plassman BL, et al. Factors associated with cognitive evaluations in the United States. Neurology 2015; 84: 64–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Chodosh J, Petitti DB, Elliott M, et al. Physician recognition of cognitive impairment: evaluating the need for improvement. J Am Geriatr Soc 2004; 52: 1051–59. [DOI] [PubMed] [Google Scholar]
  • 48.Mitchell AJ, Meader N, Pentzek M. Clinical recognition of dementia and cognitive impairment in primary care: a meta-analysis of physician accuracy. Acta Psychiatr Scand 2011; 124: 165–83. [DOI] [PubMed] [Google Scholar]
  • 49.Savva GM, Arthur A. Who has undiagnosed dementia? A cross-sectional analysis of participants of the Aging, Demographics and Memory Study. Age Ageing 2015; 44: 642–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Rosso RJ. Congressional Medicare expense report 2023. https://crsreports.congress.gov/product/pdf/IF/IF1083 (accessed Oct 9, 2023).
  • 51.American College of Surgeons. ACS NSQIP surgical risk calculator. https://riskcalculator.facs.org/RiskCalculator/ (accessed Nov 6, 2020). [PubMed]
  • 52.Stacey D, Légaré F, Lewis K, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev 2017; 4: CD001431. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1942264-supplement-1.pdf (746.9KB, pdf)

RESOURCES