Advertisement
American Journal of Kidney Diseases

Missed Hemodialysis Treatments: International Variation, Predictors, and Outcomes in the Dialysis Outcomes and Practice Patterns Study (DOPPS)

Open AccessPublished:August 23, 2018DOI:https://doi.org/10.1053/j.ajkd.2018.04.019

      Rationale & Objective

      Missed hemodialysis (HD) treatments not due to hospitalization have been associated with poor clinical outcomes and related in part to treatment nonadherence. Using data from the Dialysis Outcomes and Practice Patterns Study (DOPPS) phase 5 (2012-2015), we report findings from an international investigation of missed treatments among patients prescribed thrice-weekly HD.

      Study Design

      Prospective observational study.

      Setting & Participants

      8,501 patients participating in DOPPS, on HD therapy for more than 120 days, from 20 countries. Longitudinal and cross-sectional analyses were performed based on the 4,493 patients from countries in which 4-month missed treatment risk was > 5%.

      Predictors

      The main predictor of patient outcomes was 1 or more missed treatments in the 4 months before DOPPS phase 5 enrollment; predictors of missed treatments included country, patient characteristics, and clinical factors.

      Outcomes

      Mortality, hospitalization, laboratory measures, patient-reported outcomes, and 4-month missed treatment risk.

      Analytical Approach

      Outcomes were assessed using Cox proportional hazards, logistic, and linear regression, adjusting for case-mix and country.

      Results

      The 4-month missed treatment risk varied more than 50-fold across all 20 DOPPS countries, ranging from < 1% in Italy and Japan to 24% in the United States. Missed treatments were more likely with younger age, less time on dialysis therapy, shorter HD treatment time, lower Kt/V, longer travel time to HD centers, and more symptoms of depression. Missed treatments were positively associated with all-cause mortality (HR, 1.68; 95% CI, 1.37-2.05), cardiovascular mortality, sudden death/cardiac arrest, hospitalization, serum phosphorus level > 5.5 mg/dL, parathyroid hormone level > 300 pg/mL, hemoglobin level < 10 g/dL, higher kidney disease burden, and worse general and mental health.

      Limitations

      Possible residual confounding; temporal ambiguity in the cross-sectional analyses.

      Conclusions

      In the countries with a 4-month missed treatment risk > 5%, HD patients were more likely to die, be hospitalized, and have poorer patient-reported outcomes and laboratory measures when 1 or more missed treatments occurred in a 4-month period. The large variation in missed treatments across 20 nations suggests that their occurrence is potentially modifiable, especially in the United States and other countries in which missed treatment risk is high.

      Visual Abstract

      Index Words

      Editorial, p. 625
      Hemodialysis (HD) for treatment of end-stage kidney disease is increasing worldwide and is complex and time consuming for patients.
      • Nissenson A.R.
      • Maddux F.W.
      • Velez R.L.
      • Mayne T.J.
      • Parks J.
      Accountable care organizations and ESRD: the time has come.
      • Ortiz A.
      • Covic A.
      • Fliser D.
      • et al.
      Epidemiology, contributors to, and clinical trials of mortality risk in chronic kidney failure.
      • Robinson B.M.
      • Akizawa T.
      • Jager K.J.
      • Kerr P.G.
      • Saran R.
      • Pisoni R.L.
      Factors affecting outcomes in patients reaching end-stage kidney disease worldwide: differences in access to renal replacement therapy, modality use, and haemodialysis practices.
      Typically, HD is provided thrice weekly for several hours each treatment, which imposes considerable burdens on patients. HD patients vary greatly in the extent of being bothered by the burdens of kidney disease and HD therapy that may affect patient adherence to their treatment regimens.
      • Mapes D.L.
      • Bragg-Gresham J.L.
      • Bommer J.
      • et al.
      Health-related quality of life in the Dialysis Outcomes and Practice Patterns Study (DOPPS).
      • Rayner H.C.
      • Zepel L.
      • Fuller D.S.
      • et al.
      Recovery time, quality of life, and mortality in hemodialysis patients: the Dialysis Outcomes and Practice Patterns Study (DOPPS).
      • Lopes A.A.
      • Bragg-Gresham J.L.
      • Elder S.J.
      • et al.
      Independent and joint associations of nutritional status indicators with mortality risk among chronic hemodialysis patients in the Dialysis Outcomes and Practice Patterns Study (DOPPS).
      • Saran R.
      • Bragg-Gresham J.L.
      • Rayner H.C.
      • et al.
      Nonadherence in hemodialysis: associations with mortality, hospitalization, and practice patterns in the DOPPS.
      • Hecking M.
      • Karaboyas A.
      • Saran R.
      • et al.
      Dialysate sodium concentration and the association with interdialytic weight gain, hospitalization, and mortality.
      • Kugler C.
      • Maeding I.
      • Russell C.L.
      Non-adherence in patients on chronic hemodialysis: an international comparison study.
      • Martins M.T.
      • Silva L.F.
      • Kraychete A.
      • et al.
      Potentially modifiable factors associated with non-adherence to phosphate binder use in patients on hemodialysis.
      Missing a prescribed HD treatment (not due to hospitalization) can be considered as one form of nonadherence. In 1999, Bleyer et al
      • Bleyer A.J.
      • Hylander B.
      • Sudo H.
      • et al.
      An international study of patient compliance with hemodialysis.
      noted a much higher occurrence of missed treatments among HD patients in 4 US dialysis centers compared with 4 Japanese and 1 Swedish center. Subsequently, missed treatments have been described more extensively and shown to be associated with higher mortality and hospitalization rates.
      • Saran R.
      • Bragg-Gresham J.L.
      • Rayner H.C.
      • et al.
      Nonadherence in hemodialysis: associations with mortality, hospitalization, and practice patterns in the DOPPS.
      • Bleyer A.J.
      • Hylander B.
      • Sudo H.
      • et al.
      An international study of patient compliance with hemodialysis.
      • Leggat Jr., J.E.
      Adherence with dialysis: a focus on mortality risk.
      • Lacson Jr., E.
      • Wang W.
      • Lazarus J.M.
      • Hakim R.M.
      Hemodialysis facility-based quality-of-care indicators and facility-specific patient outcomes.
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      • Leggat Jr., J.E.
      • Orzol S.M.
      • Hulbert-Shearon T.E.
      • et al.
      Noncompliance in hemodialysis: predictors and survival analysis.
      • Obialo C.I.
      • Hunt W.C.
      • Bashir K.
      • Zager P.G.
      Relationship of missed and shortened hemodialysis treatments to hospitalization and mortality: observations from a US dialysis network.
      Numerous factors have been shown to be related to missed treatments in HD patients, not all of which may be due to nonadherence per se, including logistic factors (eg, unreliable transportation or long travel time to HD unit, inclement weather, and whether HD session was scheduled for Saturdays or holidays), effect of active medical issues (eg, gastrointestinal upset, vascular access dysfunction, alcohol/drug abuse, chronic pain, and depression), or patient cultural, religious, or demographic factors (eg, limited health literacy, age, race/ethnicity, or marital support/status).
      Missing treatments is potentially modifiable and, if the rate of missed treatments were reduced, could lead to improved patient outcomes. The Dialysis Outcomes and Practice Patterns Study (DOPPS) is an international prospective cohort study of HD care and outcomes and provides a unique opportunity for evaluation of missed treatments across a diverse set of international health care delivery/financing systems.
      • Dor A.
      • Pauly M.V.
      • Eichleay M.A.
      • Held P.J.
      End-stage renal disease and economic incentives: the International Study of Health Care Organization and Financing (ISHCOF).
      The present international study across 20 countries using DOPPS phase 5 (2012-2015) data evaluated the phenomenon of missed treatments in HD patients and its relationships with patient characteristics, treatment measures, and clinical and patient-reported outcomes (PROs).

      Methods

      Patients and Data Collection

      Patients were randomly selected to participate in the DOPPS from randomly selected HD facilities within each country to provide national HD patient samples as described previously.
      • Pisoni R.L.
      • Gillespie B.W.
      • Dickinson D.M.
      • Chen K.
      • Kutner M.H.
      • Wolfe R.A.
      The Dialysis Outcomes and Practice Patterns Study (DOPPS): design, data elements, and methodology.
      • Young E.W.
      • Goodkin D.A.
      • Mapes D.L.
      • et al.
      The Dialysis Outcomes and Practice Patterns Study: an international hemodialysis study.
      • Pisoni R.L.
      • Bieber B.A.
      • Al Wakeel J.
      • et al.
      The Dialysis Outcomes and Practice Patterns Study phase 5 in the Gulf Cooperation Council countries: design and study methods.
      Two prior publications
      • Rayner H.C.
      • Greenwood R.
      • MacTier R.
      • et al.
      Estimated life expectancy of UK HD patients if clinical practice guidelines are met.
      • Robinson B.M.
      • Fuller D.S.
      • Dykstra D.M.
      • et al.
      The DOPPS practice monitor: rationale and methods for an initiative to monitor the new US bundled dialysis payment system.
      have shown the DOPPS sampling design to yield findings closely corresponding with national data from the United States and United Kingdom. Study approval was obtained from a central institutional review board and as required by national and local ethics committees. Informed consent was obtained from all study patients. DOPPS 5 data were collected from March 2012 to July 2015 from 11,488 HD patients 18 years or older at 439 HD facilities in Australia, Belgium, Canada, China, the 6 Gulf Cooperation Council (GCC) countries (Bahrain, Qatar, Kuwait, Oman, Saudi Arabia, and United Arab Emirates), Germany, Italy, Japan, New Zealand, Russia, Spain, Sweden, Turkey, United Kingdom, and United States. Further details of the study sample are shown in Figure 1, with: (1) analyses across all countries, including 8,501 patients having missed treatments data for the entire 4-month baseline period immediately before DOPPS 5 study enrollment; and (2) additional analyses restricted to countries (Australia, New Zealand, Canada, GCC, Russia, Sweden, Turkey, United Kingdom, and United States) having > 5% of patients with 1 or more missed treatments during the 4-month period (n = 4,493 patients). This 5% cutoff was predetermined to identify countries in which missed treatment frequency was sufficiently large to understand factors related to this phenomenon. In the United States, patients were not included from one large dialysis organization due to missed treatment data not being provided. Baseline demographic data, laboratory values, and medications were based on data available in each patient’s medical record at study enrollment. Comorbidity data reflect comorbid conditions as of the study enrollment date. Mortality and hospitalization events were collected during study follow-up; median (quartile 1 [Q1]-Q3) study follow-up was 1.65 (Q1-Q3, 0.86-2.44) years. Vascular access dysfunction and emergency department/urgent visits were not captured during the period used in defining missed treatments.
      Figure thumbnail gr1
      Figure 1Sample flow chart. Abbreviation: DOPPS, Dialysis Outcomes and Practice Patterns Study.
      The primary measure of interest was whether an HD patient missed any scheduled HD treatments in the 4-month period immediately before DOPPS 5 study enrollment that were not due to a hospitalization. These data were collected from each study site in answering the question, “How many dialysis treatments did the patient miss? (Do not count time in the hospital as missed),” with response categories of 0, 1, 2 to 3, and 4 or more missed treatments each month. Missed treatments were to be reported by study sites based on each patient’s medical record.
      A patient questionnaire was completed by most patients shortly after the 4-month period, when missed treatment information was collected (after a median of 18 [Q1-Q3, 8-37] days). From the patient questionnaire, the following PRO measures were calculated: perceived general health, burden of chronic kidney disease, Kidney Disease Quality of Life Physical Component Summary (PCS) and Mental Component Summary (MCS) scores (measured on a 0-100 scale, with a higher score indicating better health), and a 10-item Center for Epidemiologic Studies Depression (CES-D) scale assessing self-reported symptoms of depression, scored 0-30, with a score ≥10 indicating symptoms of possible depression.
      • Lopes A.A.
      • Albert J.M.
      • Young E.W.
      • et al.
      Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.
      Patients self-reported on how long it typically takes to travel to their dialysis facility, which was dichotomized as 1 or less hour and more than 1 hour.

      Data Analysis

      Predictors of a patient missing 1 or more HD treatments during 4 months were analyzed using logistic regression. Associations with outcomes were estimated using: (1) linear regression for PROs treated as continuous outcomes, (2) logistic regression for laboratory measures treated as binary outcomes, and (3) Cox regression for time to death and first hospitalization while accounting for facility clustering using the robust sandwich variance estimator.
      • Klein J.
      • Moeschberger M.
      Survival Analysis: Techniques for Censored and Truncated Data.
      Linear and logistic regression analyses were performed using generalized estimating equations that accounted for facility clustering effects. Analyses were adjusted for potential confounders, including 13 comorbid conditions, as indicated in the tables. For mortality and hospitalization analyses, time at risk started at study enrollment (corresponding to the end of the 4-month missed treatment baseline period) and ended with the earliest of the following events: date of death (for mortality analyses), date of first hospitalization event (for hospitalization analyses), end of follow-up, or 7 days after leaving the facility due to transfer to another HD facility, modality switch, recovery of kidney function, or kidney transplantation. The probability of a type 1 error may exceed 0.05 when many comparisons are tested. To protect against this, we selected as predictors of an outcome only variables hypothesized a priori by study investigators or previously shown to be associated with that outcome. However, we did not attempt to “correct” for multiple comparisons, using a statistical significance correction procedure such as the Bonferroni or Benjamini-Hochberg method.
      • Bonferroni C.E.
      Teoria statistica delle classi e calcolo delle probabilità.
      • Benjamini Y.
      • Hochberg Y.
      Controlling the false discovery rate: a practical and powerful approach to multiple testing.
      Missing data were imputed using the sequential regression multiple imputation method implemented by IVEware
      • Raghunathan T.E.
      • Solenberger P.W.
      • Van Hoewyk J.
      IVEware: Imputation and Variance Estimation Software.
      and analyzed using the MIAnalyze procedure in SAS/STAT, version 9.4. All analyses used SAS software, version 9.4 (SAS Institute).

      Results

      Information regarding missed scheduled HD treatments during the 4-month baseline period was reported for 87% (8,501) of 9,731 eligible patients across the 20 study countries. The proportion of patients missing at least 1 treatment in the 4-month baseline period varied from 20% to 24% in the GCC and United States to < 1% in Italy and Japan (Table 1; Fig S1). Most of these patients missed 1 or more treatments during only 1 month in the 4-month period (Table 1). In many countries, the percentage of patients who missed 1 or more treatments over 4 months varied greatly across facilities (Fig 2). In data not shown, no consistent pattern was observed between season of the year and missed treatment occurrence, either overall or within the United States.
      Table 1Distribution of Missed Sessions by Country
      CountryNo. of PtsNo. of FacilitiesNo. of mo Pt Missed ≥ 1 Session/mo (Percentage of Pts)
      Average over 4-month period.
      Pts Missing ≥ 1 Session in 4 moMonthly % of Pts Missing ≥ 1 Session
      Average over 4-month period.
      01234
      US1,8686976.1%11.3%4.5%4.0%4.1%23.9%12.2%
      GCC6453680.0%10.4%3.9%1.4%4.3%20.0%9.9%
      Russia3912086.5%6.7%2.3%1.8%2.8%13.6%7.0%
      ANZ3531889.8%6.2%1.4%1.1%1.4%10.2%4.5%
      Canada3322090.4%5.1%1.2%1.8%1.5%9.6%4.7%
      UK3212091.0%7.8%0.9%0.3%0.0%9.0%2.6%
      Sweden2891893.1%4.2%0.7%0.7%1.4%6.9%3.3%
      Turkey2941594.2%4.4%0.7%0.0%0.7%5.8%2.1%
      Germany4932297.2%2.0%0.8%0.0%0.0%2.8%0.9%
      Spain5082297.6%2.0%0.4%0.0%0.0%2.4%0.7%
      Belgium3941997.7%2.0%0.3%0.0%0.0%2.3%0.6%
      China7954597.9%1.4%0.8%0.0%0.0%2.1%0.7%
      Italy3991899.3%0.5%0.3%0.0%0.0%0.8%0.3%
      Japan1,4195899.7%0.1%0.0%0.0%0.2%0.4%0.2%
      Note: Sample restricted to patients on dialysis therapy longer than 120 days and receiving thrice-weekly HD, with 4 months of attendance information. All other analyses shown in this article were restricted to countries in which ≥ 5% of patients missed 1 or more HD session in 4 months.
      Abbreviations: ANZ, Australia-New Zealand; GCC, Gulf Cooperation Council (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and United Arab Emirates); HD, hemodialysis; pt, patient; UK, United Kingdom; US, United States.
      a Average over 4-month period.
      Figure thumbnail gr2
      Figure 2Missed hemodialysis (HD) treatments: facility distribution by country. Percentage of patients who missed at least 1 HD session over the last 4 months; Dialysis Outcomes and Practice Patterns Study (DOPPS) phase 5 (2012-2015) patients on dialysis therapy longer than 120 days, receiving thrice-weekly HD, restricted to facilities with more than 6 qualifying patients; ◊ represents the facility mean. Abbreviations: ANZ, Australia-New Zealand; Bel, Belgium; Can, Canada; Chi, China; Fac, facilities; GCC, Gulf Cooperation Council; Ger, Germany; Ita, Italy; Jap, Japan; Rus, Russia; Spa, Spain; Swe, Sweden; Tur, Turkey; UK, United Kingdom; US, United States.
      The remaining analyses, describing patient characteristics, predictors, and outcomes associated with missing at least 1 HD session in the 4-month baseline period, were restricted to countries in which 4-month risk for missed treatments exceeded 5% (4,493 patients). Table 2 shows unadjusted comparisons of characteristics for patients who missed versus did not miss at least 1 treatment. Additional analyses revealed no consistent pattern in characteristics of patients for whom missed treatment information was reported versus not reported (Table S1).
      Table 2Characteristics of Patients Who Missed Versus Did Not Miss 1 or More HD Treatments Over 4 Months
      No. of Missed Treatments
      0≥ 1
      No. of patients3,731762
      Age, y60.7 ± 15.156.0 ± 15.2
      Dialysis vintage, y4.8 ± 4.94.1 ± 4.0
      Body mass index, kg/m227.4 ± 6.527.8 ± 6.9
      Postdialysis weight, kg76.1 ± 20.278.5 ± 21.2
      Male sex57%58%
      Single-pool Kt/V1.6 ± 0.31.5 ± 0.3
      Blood flow, prescribed, mL/min367.8 ± 74.9381.7 ± 83.3
      Treatment time, min240 [210-241]220 [196-240]
      Urine output ≥ 200 mL/d28%31%
      Lives alone17%18%
      Lives in assisted/nursing home3%2%
      Employed (full- or part-time)14%14%
      Study follow-up time, y1.66 [0.94-2.49]1.38 [0.77-2.34]
      Laboratory measurements
       Hemoglobin, g/dL11.2 ± 1.310.8 ± 1.4
       Serum albumin, g/dL3.8 ± 0.53.8 ± 0.5
       Serum creatinine, mg/dL8.4 ± 2.78.9 ± 3.2
       Serum potassium, mEq/L4.9 ± 0.74.8 ± 0.8
       Serum phosphorus, mg/dL5.1 ± 1.65.6 ± 2.0
       PTH, pg/mL295 [163-518]336 [182-587]
       Hemoglobin < 10 g/dL34%44%
       Serum creatinine < 7.5 mg/dL49%56%
       Serum phosphorus > 5.5 mg/dL16%22%
       PTH > 300 pg/mL39%37%
       IDWL > 4.2%20%25%
      Comorbid conditions
       Diabetes46%53%
       Coronary artery disease33%31%
       Congestive heart failure23%27%
       CVA disease13%10%
       Peripheral vascular disease21%22%
       Hypertension85%86%
       Other CV disease22%18%
       Recurrent cellulitis, gangrene7%9%
       Cancer other than skin10%6%
       Gastrointestinal bleeding4%5%
       Psychiatric disorder10%12%
       Lung disease11%11%
       Neurologic disease10%9%
      Patient questionnaire
       Completed questionnaire
      Excludes 715 patients who did not have the opportunity to complete the patient questionnaire per study protocol, resulting in 3,210 and 568 PQs administered to patients in the missed treatments = 0 vs missed treatments ≥1 groups, respectively.
      66% (n = 2,121)59% (n = 336)
       Travel time to facility ≥1 h (%)6%12%
       CES-D score8.7 ± 5.411.0 ± 6.2
       CES-D > 1034%49%
       PCS score (SF-12)
      PCS and MCS scores have a range of 0 to 100 and were designed to have a mean ± standard deviation score of 50±10 in a representative sample of the US population (not ESRD specific). Higher scores indicate better health status.
      36.0 ± 10.435.3 ± 11.1
       MCS score (SF-12)
      PCS and MCS scores have a range of 0 to 100 and were designed to have a mean ± standard deviation score of 50±10 in a representative sample of the US population (not ESRD specific). Higher scores indicate better health status.
      45.0 ± 11.741.9 ± 11.5
       Perceived burden of kidney disease40.9 ± 28.136.9 ± 26.8
       Perceived general health44.6 ± 27.638.7 ± 27.8
       Recovery time < 2 h31%30%
       Transplant waitlisted24%25%
      Note: Data shown as mean ± standard deviation, percent, or median [quartile 1-quartile 3]. Restricted to patients with 4 months of information about missed sessions, dialysis vintage longer than 120 days, and receiving thrice-weekly HD; also restricted to countries with >5% of patients missing 1 or more sessions in 4 months (United States, GCC, Russia, Australia-New Zealand, Canada, United Kingdom, Sweden, and Turkey). All variables were missing in <5% of sample, except for single-pool Kt/V (missing in 18%) and PTH level (missing in 10%); residual urine volume was missing in 9% of sample (once US large dialysis organizations without the measure are excluded), and patient questionnaire responses varied from 23% missing for composite variables MCS and PCS scores to 2% missing responses to questions about burden of kidney disease; question about travel time was not collected in Russia and Turkey.
      Abbreviations: CES-D, Center for Epidemiologic Studies Depression; CV, cardiovascular; CVA, cerebrovascular accident; GCC, Gulf Cooperation Council (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and United Arab Emirates); HD, hemodialysis; IDWL, intradialytic weight loss; MCS, Mental Component Summary; PCS, Physical Component Summary; PTH, parathyroid hormone; SF-12, 12-Item Short Form.
      a Excludes 715 patients who did not have the opportunity to complete the patient questionnaire per study protocol, resulting in 3,210 and 568 PQs administered to patients in the missed treatments = 0 vs missed treatments ≥1 groups, respectively.
      b PCS and MCS scores have a range of 0 to 100 and were designed to have a mean ± standard deviation score of 50 ± 10 in a representative sample of the US population (not ESRD specific). Higher scores indicate better health status.
      Logistic regression analyses, adjusted for case-mix and country, were used to investigate the relationship of missing 1 or more treatments over 4 months with various factors (Table 3). Patients were more likely to have 1 or more missed treatments if younger or having shorter dialysis vintage, shorter prescribed treatment time, lower achieved Kt/V, depression symptoms (CES-D score > 10), or typically traveled more than 1 hour to HD sessions. The relation with travel time was more pronounced in the United States (adjusted odds ratio [aOR], 3.17; 95% confidence interval [CI], 1.41-7.16) compared with other countries (aOR, 1.60; 95% CI, 0.94-2.72). Patient comorbid factors, or having residual renal output > 200 mL of urine per day, were not strongly associated with odds of at least 1 missed treatment. Furthermore, the aOR of 1 or more missed treatments differed little by sex in all countries except the GCC, where the odds were nominally greater in males, though this was not statistically significant (aOR, 1.44; 95% CI, 0.97-2.13; P = 0.07).
      Table 3aOR of Missing Versus Not Missing 1 or More HD Treatments Over 4 Months According to Patient and Treatment Characteristics
      PredictoraOR (95% CI)P
      Age, per 10 y older0.83 (0.78-0.88)<0.001
      Sex, male vs female1.10 (0.94-1.28)0.2
      Dialysis vintage, per 1 y longer0.97 (0.95-0.99)0.002
      Travel time to facility >1 h
      Travel time and CES-D measures were available only for patients who completed the patient questionnaire (n available = 2,457); travel time was not available for patients in Russia and Turkey.
      1.98 (1.30-3.02)0.001
      Depression symptoms CES-D >10
      Travel time and CES-D measures were available only for patients who completed the patient questionnaire (n available = 2,457); travel time was not available for patients in Russia and Turkey.
      1.51 (1.16-1.96)0.003
      GI bleeding, yes vs no1.48 (0.98-2.23)0.06
      Single-pool Kt/V, per 0.1 greater0.96 (0.93-0.99)0.01
      Treatment time, per 10 min longer0.90 (0.87-0.93)<0.001
      Urine output ≥ 200 mL/d1.20 (0.90-1.60)0.2
      Lives alone
      Lives alone is compared to combinations of other arrangements: with family or friends or assisted/nursing home.
      1.39 (1.05-1.83)0.02
      Employed (part- or fulltime)0.85 (0.61-1.19)0.4
      Note: Outcome is aOR for missing 1 or more treatments (vs missing 0 treatments) over 4 months, showing association with other measures in a logistic model, adjusting for country, age, sex, vintage, and 13 comorbid conditions (diabetes mellitus, cerebrovascular accident, congestive heart failure, peripheral vascular disease, coronary heart disease, hypertension, other cardiovascular diseases, recurrent cellulitis/gangrene, cancer other than skin, GI bleeding, lung disease, neurologic disease, and psychiatric disorder). Variables having P > 0.3 that were not displayed were cancer, coronary artery disease, cardiovascular disease, congestive heart failure, diabetes, hypertension, lung disease, neurologic disease, and recurrent cellulitis; peripheral vascular disease had P = 0.20 and cerebrovascular disease had P = 0.18. Restricted to countries with > 5% of patients missing an HD treatment in 4 months (United States, GCC, Russia, Australia-New Zealand, Canada, United Kingdom, Sweden, and Turkey).
      Abbreviations: aOR, adjusted odds ratio; CES-D, Center for Epidemiologic Studies Depression; CI, confidence interval; GCC, Gulf Cooperation Council; GI, gastrointestinal; HD, hemodialysis.
      a Travel time and CES-D measures were available only for patients who completed the patient questionnaire (n available = 2,457); travel time was not available for patients in Russia and Turkey.
      b Lives alone is compared to combinations of other arrangements: with family or friends or assisted/nursing home.
      Associations were examined between having 1 or more missed treatments and PROs, levels of selected laboratory values, mortality, and hospitalization in models using extensive adjustments for patient case-mix and country. Patients having 1 or more missed treatments were more likely to have hyperphosphatemia (serum phosphorus > 5.5 mg/dL), parathyroid hormone (PTH) level > 300 pg/mL, hemoglobin level < 10 g/dL, and single-pool Kt/V < 1.2 (Table 4). However, at the cutpoints shown, values for serum potassium, albumin, creatinine, and intradialytic weight loss differed little for patients with versus without 1 or more missed treatments over 4 months. In analyses not shown, differences were small between patients with versus without 1 or more missed treatments when outcomes of serum potassium, PTH, and creatinine levels were analyzed as continuous variables.
      Table 4aOR of Laboratory Outcomes for Patients Missing Versus Not Missing 1 or More HD Treatments Over 4 Months
      Laboratory OutcomeaOR (95% CI)P
      Serum phosphorus > 5.5 mg/dL1.24 (1.01-1.53)0.04
      PTH > 300 pg/mL1.29 (1.09-1.52)0.003
      PTH > 600 pg/mL1.20 (0.99-1.44)0.06
      Hemoglobin < 10 g/dL1.43 (1.16-1.75)0.001
      Single-pool Kt/V < 1.21.60 (1.15-2.23)0.007
      Serum albumin < 3.0 g/dL1.28 (0.88-1.86)0.2
      Serum creatinine < 7.5 mg/dL1.24 (1.01-1.52)0.04
      Serum potassium > 5.9 mEq/L1.07 (0.77-1.49)0.70
      IDWL > 4.2%
      IDWL > 4.2% corresponds to the top 25th percentile of the IDWL distribution in the study sample.
      1.10 (0.89-1.37)0.4
      Note: Association of missing 1 or more treatment in a 4-month period as a predictor of each of these outcome measures in a logistic model, with additional adjustments for country, age, vintage, sex, and 13 comorbid conditions (listed in Table 3 footnote). Restricted to countries with > 5% patients missing an HD treatment in 4 months (United States, GCC, Russia, Australia-New Zealand, Canada, United Kingdom, Sweden, and Turkey).
      Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; GCC, Gulf Cooperation Council; HD, hemodialysis; IDWL, intradialytic weight loss; PTH, parathyroid hormone.
      a IDWL > 4.2% corresponds to the top 25th percentile of the IDWL distribution in the study sample.
      Five PROs were examined as continuous variables to determine to what extent these PROs differed for patients with versus without 1 or more missed treatments over 4 months (Table 5). Patients with 1 or more missed treatments reported substantially greater burden of kidney disease, higher mean CES-D scores (indicative of greater depression symptoms), poorer perceived general health, and lower MCS scores. Little difference was seen in PCS scores for patients with versus without 1 or more missed treatments.
      Table 5Patient-Reported Outcomes for Patients Missing Versus Not Missing 1 or More HD Treatments Over 4 Months
      Patient-Reported OutcomeEstimate (95% CI)P
      Perceived burden of kidney disease
      Higher scores indicate lower perceived disease burden and better perceived general health.
      −4.53 (−7.86 to −1.21)0.008
      Perceived general health
      Higher scores indicate lower perceived disease burden and better perceived general health.
      −6.17 (−9.47 to −2.87)<0.001
      PCS score
      PCS and MCS scores have a range of 0 to 100 and were designed to have a mean (standard deviation) score of 50 (10) in a representative sample of the US population. Higher scores indicate better health status.
      −0.81 (−2.16 to 0.54)0.2
      MCS score
      PCS and MCS scores have a range of 0 to 100 and were designed to have a mean (standard deviation) score of 50 (10) in a representative sample of the US population. Higher scores indicate better health status.
      −2.52 (−4.08 to −0.95)0.002
      CES-D score
      CES-D, self-test score is sum of 10 questions, with range of 0 to 30 (score ≥ 10 points is considered depressed).
      1.69 (0.97 to 2.41)<0.001
      Note: Associations of missing at least 1 HD treatment (vs 0) during the 4 months before DOPPS enrollment, with listed outcomes from linear regression models also adjusted for country, age, vintage, sex, and 13 comorbid conditions (listed in Table 3 footnote). Restricted to countries with > 5% patients missing an HD treatment in 4 months (United States, GCC, Russia, Australia-New Zealand, Canada, United Kingdom, Sweden, and Turkey).
      Abbreviations: CES-D, Center for Epidemiologic Studies Depression Scale; CI, confidence interval; DOPPS, Dialysis Outcomes and Practice Patterns Study; GCC, Gulf Cooperation Council; HD, hemodialysis; MCS, Mental Component Summary; PCS, Physical Component Summary.
      a Higher scores indicate lower perceived disease burden and better perceived general health.
      b PCS and MCS scores have a range of 0 to 100 and were designed to have a mean (standard deviation) score of 50 (10) in a representative sample of the US population. Higher scores indicate better health status.
      c CES-D, self-test score is sum of 10 questions, with range of 0 to 30 (score ≥ 10 points is considered depressed).
      Cox regression revealed that all-cause mortality was 68% higher for patients with versus without 1 or more missed treatments over 4 months in adjusted survival analyses (Table 6). This mortality finding was consistent when tested separately in each of 4 countries/DOPPS regions (as defined by a country’s 4-month missed treatment risk), with the following hazard ratios (HRs) of death observed: United States, 1.57 (95% CI, 1.22-2.01); GCC, 2.03 (95% CI, 1.25-3.31); Russia, 3.80 (95% CI, 2.19-6.59); and in aggregate, Australia–New Zealand, Canada, United Kingdom, Sweden, and Turkey (1.48; 95% CI, 1.02-2.13). The overall mortality association was nearly unchanged when models were additionally adjusted for serum phosphorus, potassium, and intradialytic weight loss values. Similar findings were obtained in sensitivity analyses, in which missed treatments were based on those counted in the 4 months immediately following study enrollment, which did not require survival through the entire 4-month period. The relationship of 1 or more missed treatments over 4 months with all-cause mortality did not appear to substantially differ by particular patient characteristics when the following interactions were tested (P > 0.10 in each case): age 65 years or older versus younger than 65 years (P = 0.8), female versus male sex (P = 0.2), coronary artery disease (P = 0.8), other cardiovascular disease (P = 0.9), and having an achieved single-pool Kt/V < 1.2 (P = 0.9).
      Table 6Subsequent Mortality and Hospitalization Outcomes for Patients Who Missed Versus Did Not Miss 1 or More HD Treatments Over 4-Months
      ModelModel Results: HR (95% CI)Unadjusted Rates, per 100 patient-y
      Not Adjusted
      Not adjusted models were stratified by country only.
      Adjusted
      Adjusted models were stratified by country and adjusted for age, sex, vintage, patient weight, and 13 comorbid conditions (listed in Table 3 footnote).
      No Missed Treatment≥ 1 Missed Treatment
      All-cause mortality1.53 (1.25-1.86)1.68 (1.37-2.05)14.2 (n = 765 events)22.1 (n = 190 events)
      CV-related mortality
      CV mortality includes the sudden death/cardiac arrest cases.
      1.63 (1.18-2.25)1.76 (1.25-2.48)4.3 (n = 229 events)6.9 (n = 59 events)
      Sudden death/cardiac arrest mortality1.98 (1.33-2.96)2.16 (1.39-3.36)2.4 (n = 18 events)5.1 (n = 4 events)
      CVA-related mortality1.73 (0.59-5.04)1.66 (0.55-4.95)0.3 (n = 66 events)0.5 (n = 15 events)
      CVA event (mortality or hospitalization)
      CVA events (either hospitalization or death) were limited to occurrence of a stroke.
      1.41 (0.85-2.35)1.49 (0.90-2.49)1.5 (n = 28 events)1.5 (n = 5 events)
      Infection-related mortality1.61 (0.88-2.94)1.66 (0.93-2.97)1.2 (n = 115 events)1.7 (n = 38 events)
      All-cause hospitalization1.26 (1.04-1.52)1.28 (1.04-1.57)56.1 (n = 2,058 events)70.4 (n = 466 events)
      Note: Cause-specific mortality analyses exclude US large dialysis organizations due to data-reporting issue; mortality cause categories are not mutually exclusive. Restricted to countries with > 5% patients missing an HD treatment in 4 months (United States, GCC, Russia, Australia-New Zealand, Canada, United Kingdom, Sweden, and Turkey). Country/regional variation in HRs for all-cause mortality in patients with versus without 1 or more missed treatments in 4 months, in adjusted analyses, was as follows: United States, 1.57 (95% CI, 1.22-2.01); GCC, 2.03 (95% CI, 1.25-3.31); Russia, 3.80 (95% CI, 2.19-6.59); and Australia-New Zealand, Canada, United Kingdom, Sweden, and Turkey in aggregate, 1.48 (95% CI, 1.02-2.13). Patient study follow-up time median was 1.66 [Q1-Q3, 0.94-2.49] years for patients without a missed treatment and 1.38 [Q1-Q3, 0.77-2.34] years for patients with 1 or more missed treatments.
      Abbreviations: CI, confidence interval; CV, cardiovascular; CVA, cerebrovascular accident; GCC, Gulf Cooperation Council; HD, hemodialysis; HR, hazard ratio; Q, quartile.
      a Not adjusted models were stratified by country only.
      b Adjusted models were stratified by country and adjusted for age, sex, vintage, patient weight, and 13 comorbid conditions (listed in Table 3 footnote).
      c CV mortality includes the sudden death/cardiac arrest cases.
      d CVA events (either hospitalization or death) were limited to occurrence of a stroke.
      Compared with patients not missing sessions, HRs for all-cause mortality were 1.58 (95% CI, 1.25-2.00) and 1.84 (95% CI, 1.34-2.52), respectively, for patients who had 1 or more missed treatments during 1 month versus 2 or more months of the 4-month baseline period. This result suggests a higher mortality rate with increasing number of months with missed treatments. Our sample size was inadequate to test this hypothesis in any greater detail. A similar pattern was seen when investigating the relationship of total number of missed treatments during the 4 months with mortality: HRs for death of 1.56 (95% CI, 1.21-2.03) and 1.80 (95% CI, 1.35-2.38), respectively, for patients having 1 versus 2 or more missed treatments during the 4-month baseline period, compared with no missed treatments.
      Analyses of the relationship between cause-specific mortality and 1 or more missed treatments indicated 1.76- to 2.16-fold higher rates of cardiovascular mortality and sudden death/cardiac arrest mortality for patients with versus without 1 or more missed treatments (Table 6). Nominally higher rates of cerebrovascular accident–related mortality and infection-related mortality were observed for patients with 1 or more missed treatments, but with a wide CI. The HR of cerebrovascular accident events resulting in hospitalization or death was 1.49-fold (95% CI, 0.90-2.49) higher for patients with 1 or more missed treatments over 4 months. Analyses of all-cause hospitalization indicated a 1.3-fold higher rate of hospitalization during study follow-up for patients with versus without 1 or more missed treatments over 4 months.

      Discussion

      The current study is based on data collected from 2012 to 2015 at 439 facilities in 20 countries. The 4-month risk for at least 1 missed treatment was highest in the United States at 24%, followed by 20% and 14%, respectively, in the GCC and Russia and 6% to 10% in Australia–New Zealand, United Kingdom, Canada, Sweden, and Turkey. Notably, 4-month risk for at least 1 missed treatment was even lower in 6 of the other DOPPS 5 countries, such as Japan and Italy, at <1%. Thus, some countries have been very successful in greatly limiting the occurrence of missed treatments for their patients. In the United States, missed treatment occurrence appears to have increased during the past 15 to 20 years, based on the estimate of 7.9% of US patients missing 1 or more HD treatments in 1 month, reported by Saran et al
      • Saran R.
      • Bragg-Gresham J.L.
      • Rayner H.C.
      • et al.
      Nonadherence in hemodialysis: associations with mortality, hospitalization, and practice patterns in the DOPPS.
      from 1996 to 2001, versus 12.2% in these US DOPPS 5 data (2012-2015). US representativeness in DOPPS 5 may be limited by missed treatment data not being provided by 1 US large dialysis organization, although study data included 1,868 patients in 69 US facilities across 31 states. For many of the 20 countries included in our analyses, missed treatment, based on national samples, has not been previously reported beyond the 7 countries described earlier by Saran et al.
      • Saran R.
      • Bragg-Gresham J.L.
      • Rayner H.C.
      • et al.
      Nonadherence in hemodialysis: associations with mortality, hospitalization, and practice patterns in the DOPPS.
      Consistent with prior studies showing higher mortality rates for HD patients who miss an HD treatment,
      • Leggat Jr., J.E.
      Adherence with dialysis: a focus on mortality risk.
      • Lacson Jr., E.
      • Wang W.
      • Lazarus J.M.
      • Hakim R.M.
      Hemodialysis facility-based quality-of-care indicators and facility-specific patient outcomes.
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      • Leggat Jr., J.E.
      • Orzol S.M.
      • Hulbert-Shearon T.E.
      • et al.
      Noncompliance in hemodialysis: predictors and survival analysis.
      • Obialo C.I.
      • Hunt W.C.
      • Bashir K.
      • Zager P.G.
      Relationship of missed and shortened hemodialysis treatments to hospitalization and mortality: observations from a US dialysis network.
      we observed on average 68% higher mortality for HD patients who missed versus did not miss 1 or more HD treatments during 4 months. This raises the question of why mortality is so sensitive to missing even 1 HD treatment session. Does missing 1 HD session compromise the electrolyte balance to result in long-term irreversible undesirable health outcomes; and/or does it reflect unmeasured indicators of poorer health, burden of kidney failure, poor quality of care, or poor patient adherence overall; or relate to unmeasured sociocultural causes also associated with poorer survival?
      One of the strongest associations with death we observed of missing 1 or more treatments over 4 months was due to sudden death/cardiac arrest. This could be explained by the excessive cardiovascular burden related to expanded extracellular volume, as well as by changes in electrolyte levels and vascular volume that may occur when exposed to longer intervals between HD treatments. Prior investigations have shown elevated risks for sudden death and cardiovascular-related mortality at the end of the 2 days off in a typical thrice-weekly HD treatment schedule.
      • Foley R.N.
      • Gilbertson D.T.
      • Murray T.
      • Collins A.J.
      Long interdialytic interval and mortality among patients receiving hemodialysis.
      • Bleyer A.J.
      • Russell G.B.
      • Satko S.G.
      Sudden and cardiac death rates in hemodialysis patients.
      • Zhang H.
      • Schaubel D.E.
      • Kalbfleisch J.D.
      • et al.
      Dialysis outcomes and analysis of practice patterns suggests the dialysis schedule affects day-of-week mortality.
      This relationship could be further exacerbated by missing an HD treatment, prolonging the number of days without dialysis beyond the 2 days off. Data were not collected in the DOPPS regarding “make-up” sessions, which may bias the estimations of missed treatment on mortality and other patient outcomes (eg, underestimating the true missed treatment–mortality relationship). In prior studies, patients who missed HD treatments displayed poorer control of anemia, bone mineral milieu, and blood pressure and more severe electrolyte imbalance, most being risk factors for cardiovascular-related morbidity and mortality.
      • Tentori F.
      • Blayney M.J.
      • Albert J.M.
      • et al.
      Mortality risk for dialysis patients with different levels of serum calcium, phosphorus, and PTH: the Dialysis Outcomes and Practice Patterns Study (DOPPS).
      • Robinson B.
      • Tong L.
      • Zhang J.
      • et al.
      Blood pressure levels and mortality risk among hemodialysis patients in the Dialysis Outcomes and Practice Patterns Study.
      • Jadoul M.
      • Thumma J.
      • Fuller D.S.
      • et al.
      Modifiable practices associated with sudden death among hemodialysis (HD) patients in the Dialysis Outcomes and Practice Patterns Study (DOPPS).
      • Wong M.M.Y.
      • McCullough K.P.
      • Bieber B.
      • et al.
      Interdialytic weight gain: trends, predictors, and associated outcomes in the international Dialysis Outcomes and Practice Patterns Study (DOPPS).
      • Karaboyas A.
      • Zee J.
      • Brunelli S.M.
      • et al.
      Dialysate potassium, serum potassium, mortality and arrhythmia events in hemodialysis: results from the DOPPS.
      • Ramirez S.P.B.
      • Kapke A.
      • Port F.
      • et al.
      Dialysis dose scaled to body surface area and size-adjusted, sex-specific patient mortalilty.
      In the current study, we observed patients with 1 or more missed treatments to be more likely to have hyperphosphatemia, higher PTH levels, low hemoglobin levels (< 10 g/dL), and single-pool Kt/V < 1.2, whereas serum potassium and intradialytic weight loss levels were similar to those seen in patients without 1 or more missed treatments.
      As another important outcome, all-cause hospitalization was higher among patients missing 1 or more treatments. Chan et al
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      showed that for nearly 45 million outpatient HD treatments at US Fresenius HD facilities from 2005 to 2009, patients who missed HD treatments were nearly 4 times more likely to be hospitalized or admitted to an intensive/critical care facility and twice as likely to visit the emergency department compared with patients not missing a treatment. Obialo et al
      • Obialo C.
      • Zager P.G.
      • Myers O.B.
      • Hunt W.C.
      Relationships of clinic size, geographic region, and race/ethnicity to the frequency of missed/shortened dialysis treatments.
      reported a large linear increase in subsequent age-adjusted hospitalization with each additional missed treatment that occurred during a study month among more than 15,000 US HD patients treated at Dialysis Clinics, Inc facilities from January 2007 to June 2008.
      In the current study, patients who missed 1 or more treatments over 4 months were younger and had shorter dialysis vintage and shorter mean HD session treatment times compared with patients without 1 or more missed treatments. Regarding the treatment time association, we speculate that patients who accept longer treatments may be more adherent to their treatment than patients receiving shorter HD treatments. Prior studies similarly have observed greater likelihood of missed treatments with younger age.
      • Saran R.
      • Bragg-Gresham J.L.
      • Rayner H.C.
      • et al.
      Nonadherence in hemodialysis: associations with mortality, hospitalization, and practice patterns in the DOPPS.
      • Bleyer A.J.
      • Hylander B.
      • Sudo H.
      • et al.
      An international study of patient compliance with hemodialysis.
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      • Leggat Jr., J.E.
      • Orzol S.M.
      • Hulbert-Shearon T.E.
      • et al.
      Noncompliance in hemodialysis: predictors and survival analysis.
      • Obialo C.I.
      • Hunt W.C.
      • Bashir K.
      • Zager P.G.
      Relationship of missed and shortened hemodialysis treatments to hospitalization and mortality: observations from a US dialysis network.
      • Obialo C.
      • Zager P.G.
      • Myers O.B.
      • Hunt W.C.
      Relationships of clinic size, geographic region, and race/ethnicity to the frequency of missed/shortened dialysis treatments.
      Younger patients may perceive themselves as physically healthier than their older counterparts and thus able to “get away with” missed treatments. Health literacy may in part contribute to the perceived misconception by HD patients that missing an HD session will not significantly affect health. Brar et al
      • Brar A.
      • Babakhani A.
      • Salifu M.O.
      • Jindal R.M.
      Evaluation of non-adherence in patients undergoing dialysis and kidney transplantation: United States transplantation practice patterns survey.
      found this to be the second most dominant reason after transportation problems for patients missing a scheduled HD treatment. It has also been suggested that missed treatments may be a subtle way for patients to express control over their health status.
      • Kutner N.G.
      • Zhang R.
      • McClellan W.M.
      • Cole S.A.
      Psychosocial predictors of non-compliance in haemodialysis and peritoneal dialysis patients.
      Our study results indicate that travel time to the HD facility of more than 1 hour is strongly associated with missed treatment. This finding is consistent with results by Brar et al.
      • Brar A.
      • Babakhani A.
      • Salifu M.O.
      • Jindal R.M.
      Evaluation of non-adherence in patients undergoing dialysis and kidney transplantation: United States transplantation practice patterns survey.
      Similarly, Chan et al
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      noted that transportation-related factors were strongly related to a greater likelihood of missed treatments in a large cohort of US HD patients dialyzed in Fresenius dialysis facilities. Patients with private transportation to dialysis had better dialysis attendance than patients relying on public transportation.
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      Facilitating transportation could help improve dialysis adherence. In Belgium, transportation to HD facilities is covered almost universally, with travel times to HD facilities typically being relatively short. In Japan, dialysis facilities are intentionally located near public transportation lines and routes often used by patients.
      Patients receiving maintenance HD experience many symptoms; depression is common and potentially treatable.
      • Lopes A.A.
      • Albert J.M.
      • Young E.W.
      • et al.
      Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.
      • Hedayati S.S.
      • Minhajuddin A.T.
      • Afshar M.
      • Toto R.D.
      • Trivedi M.H.
      • Rush A.J.
      Association between major depressive episodes in patients with chronic kidney disease and initiation of dialysis, hospitalization, or death.
      • Watnick S.
      • Kirwin P.
      • Mahnensmith R.
      • Concato J.
      The prevalence and treatment of depression among patients starting dialysis.
      • Weisbord S.D.
      • Mor M.K.
      • Sevick M.A.
      • et al.
      Associations of depressive symptoms and pain with dialysis adherence, health resource utilization, and mortality in patients receiving chronic hemodialysis.
      A strength of this study is the numerous PROs collected and investigated. Notably, patients who missed 1 or more treatments over 4 months had substantially greater burden of kidney disease, poorer perceived general health, and lower MCS scores and were 1.7-fold more likely to have a CES-D score > 10, which is indicative of symptoms of depression. Previous studies similarly have shown a strong relationship between higher rates of missed treatments with depression
      • Chan K.E.
      • Thadhani R.I.
      • Maddux F.W.
      Adherence barriers to chronic dialysis in the United States.
      • Weisbord S.D.
      • Mor M.K.
      • Sevick M.A.
      • et al.
      Associations of depressive symptoms and pain with dialysis adherence, health resource utilization, and mortality in patients receiving chronic hemodialysis.
      and low MCS scores.
      • DeOreo P.B.
      Hemodialysis patient-assessed functional health status predicts continued survival, hospitalization, and dialysis-attendance compliance.
      Several additional factors have been highlighted in prior studies as being associated with missed treatment, as summarized in the introduction, such as black race and day of the week. Although we were not able to ascribe day of the week with missed treatments reported in our data, our analyses, beyond those already described, similarly showed 1 or more missed treatment to be more likely for: (1) black HD patients (aOR, 1.92 [95% CI, 1.49-2.48] overall; aOR, 1.90 [95% CI, 1.47-2.45] in United States only), (2) patients who smoked (aOR, 1.57; 95% CI, 1.15-2.14), and (3) patients with a history of alcohol and/or drug abuse (aOR, 2.91; 95% CI, 1.42-5.95).
      We were very interested to understand what factors may be driving the observed large international differences in missed treatments and have been disappointed during our analyses in not being able to elucidate key factors explaining these differences. Mean age of HD patients tended to be younger for the 3 countries with the highest percentage of missed treatments (GCC, Russia, and United States), but some countries with considerably lower percentages of missed treatments (China and Turkey) had mean ages younger than US patients (Table S2). Reported alcohol and substance abuse were relatively low in each country and did not appear to explain the large variations in missed treatment across countries (Table S2). Thus, the key factors explaining international missed treatment differences remain to be understood. Ultimately, cultural or environmental differences may play a major role, or other unmeasured factors (eg, transportation services). It would be meaningful to understand the underlying reasons for these large international differences in missed treatments and encourage future data collection specifically designed for this purpose. Relatedly, some missed treatments may be under a patient’s behavioral decision control versus others that result from situations not necessarily within a patient’s control (eg, transportation failures and inclement weather). These 2 missed treatment types may relate differently with outcomes. Although some missed treatments may be a result of behavioral nonadherence, nonadherence is commonly observed in many different chronic conditions and can substantially vary across countries, and not necessarily in a consistent pattern across chronic conditions.
      • Sabate E.
      Adherence to Long-term Therapies: Evidence for Action.
      • Larsen J.
      • Stovring H.
      • Kragstrup J.
      • Hansen D.G.
      Can differences in medical drug compliance between European countries be explained by social factors: analyses based on data from the European Social Survey, round 2.
      • Fissell R.B.
      • Karaboyas A.
      • Bieber B.A.
      • et al.
      Phosphate binder pill burden, patient-reported non-adherence, and mineral bone disorder markers: Findings from the DOPPS.
      Our study has several limitations. Our observational study design limits causal inference due to possible residual confounding despite analyses accounting for many factors. The cross-sectional analyses do not inform the direction of associations between missed treatment and other studied variables. Although our expectation is that study sites reported missed treatments based on each patient’s medical record, we do not know if the method of ascertaining missed treatments or comorbid conditions was consistent across DOPPS study sites. Despite these limitations, our study has strong points, including: (1) precise definition of missed treatment, with examination of contributing factors, possible confounders, and numerous outcomes, including PROs, based on a priori hypotheses; and (2) a large cohort sample based on randomly selected national samples of HD facilities in 20 countries.
      The present large international study consistently demonstrates substantially poorer health-related outcomes associated with missing treatments in HD patients. These outcomes include higher rates of mortality and hospitalization; poorer anemia, serum phosphorus, and PTH control; poorer dialysis adequacy achievement; poorer MCS scores; and greater likelihood of depression symptoms and kidney disease burden. However, the >50-fold difference in 4-month missed treatment risk across countries and large variability in missed treatment occurrence across facilities in individual countries strongly suggest that its occurrence is modifiable. Across the 20 DOPPS countries, the United States displayed the highest 4-month missed treatment risk at 24%, compared with < 1% in Italy and Japan. The current study provides additional insights into specific factors related to missed treatment. Future studies are needed to understand the key factors responsible for the observed large international missed treatment differences and to what extent reducing missed treatments in HD patients will affect patient mortality, hospitalization, and quality of life.

      Acknowledgements

      Heather Van Doren, MFA, senior medical editor with Arbor Research Collaborative for Health, provided editorial assistance on this manuscript.
      Peer Review: Received October 13, 2017. Evaluated by 3 external peer reviewers, with direct editorial input from a Statistics/Methods Editor, an Associate Editor, and the Editor-in-Chief. Accepted in revised form April 26, 2018.
      Correction Notice: This article was amended on September 13 and 20, 2018 to correct statements in the abstract regarding the number of countries analyzed.

      Supplementary Material

      • Supplementary Table S1 (PDF)

        Characteristics of patients who missed versus did not miss ≥1 HD treatment over 4 months, including patients missing information about missed treatments.

      References

        • Nissenson A.R.
        • Maddux F.W.
        • Velez R.L.
        • Mayne T.J.
        • Parks J.
        Accountable care organizations and ESRD: the time has come.
        Am J Kidney Dis. 2012; 59: 724-733
        • Ortiz A.
        • Covic A.
        • Fliser D.
        • et al.
        Epidemiology, contributors to, and clinical trials of mortality risk in chronic kidney failure.
        Lancet. 2014; 383: 1831-1843
        • Robinson B.M.
        • Akizawa T.
        • Jager K.J.
        • Kerr P.G.
        • Saran R.
        • Pisoni R.L.
        Factors affecting outcomes in patients reaching end-stage kidney disease worldwide: differences in access to renal replacement therapy, modality use, and haemodialysis practices.
        Lancet. 2016; 388: 294-306
        • Mapes D.L.
        • Bragg-Gresham J.L.
        • Bommer J.
        • et al.
        Health-related quality of life in the Dialysis Outcomes and Practice Patterns Study (DOPPS).
        Am J Kidney Dis. 2004; 44: 54-60
        • Rayner H.C.
        • Zepel L.
        • Fuller D.S.
        • et al.
        Recovery time, quality of life, and mortality in hemodialysis patients: the Dialysis Outcomes and Practice Patterns Study (DOPPS).
        Am J Kidney Dis. 2014; 64: 86-94
        • Lopes A.A.
        • Bragg-Gresham J.L.
        • Elder S.J.
        • et al.
        Independent and joint associations of nutritional status indicators with mortality risk among chronic hemodialysis patients in the Dialysis Outcomes and Practice Patterns Study (DOPPS).
        J Ren Nutr. 2010; 20: 224-234
        • Saran R.
        • Bragg-Gresham J.L.
        • Rayner H.C.
        • et al.
        Nonadherence in hemodialysis: associations with mortality, hospitalization, and practice patterns in the DOPPS.
        Kidney Int. 2003; 64: 254-262
        • Hecking M.
        • Karaboyas A.
        • Saran R.
        • et al.
        Dialysate sodium concentration and the association with interdialytic weight gain, hospitalization, and mortality.
        Clin J Am Soc Nephrol. 2012; 7: 92-100
        • Kugler C.
        • Maeding I.
        • Russell C.L.
        Non-adherence in patients on chronic hemodialysis: an international comparison study.
        J Nephrol. 2011; 24: 366-375
        • Martins M.T.
        • Silva L.F.
        • Kraychete A.
        • et al.
        Potentially modifiable factors associated with non-adherence to phosphate binder use in patients on hemodialysis.
        BMC Nephrol. 2013; 14: 208
        • Bleyer A.J.
        • Hylander B.
        • Sudo H.
        • et al.
        An international study of patient compliance with hemodialysis.
        JAMA. 1999; 281: 1211-1213
        • Leggat Jr., J.E.
        Adherence with dialysis: a focus on mortality risk.
        Semin Dial. 2005; 18: 137-141
        • Lacson Jr., E.
        • Wang W.
        • Lazarus J.M.
        • Hakim R.M.
        Hemodialysis facility-based quality-of-care indicators and facility-specific patient outcomes.
        Am J Kidney Dis. 2009; 54: 490-497
        • Chan K.E.
        • Thadhani R.I.
        • Maddux F.W.
        Adherence barriers to chronic dialysis in the United States.
        J Am Soc Nephrol. 2014; 25: 2642-2648
        • Leggat Jr., J.E.
        • Orzol S.M.
        • Hulbert-Shearon T.E.
        • et al.
        Noncompliance in hemodialysis: predictors and survival analysis.
        Am J Kidney Dis. 1998; 32: 139-145
        • Obialo C.I.
        • Hunt W.C.
        • Bashir K.
        • Zager P.G.
        Relationship of missed and shortened hemodialysis treatments to hospitalization and mortality: observations from a US dialysis network.
        Clin Kidney J. 2012; 5: 315-319
        • Dor A.
        • Pauly M.V.
        • Eichleay M.A.
        • Held P.J.
        End-stage renal disease and economic incentives: the International Study of Health Care Organization and Financing (ISHCOF).
        Int J Health Care Finance Econ. 2007; 7: 73-111
        • Pisoni R.L.
        • Gillespie B.W.
        • Dickinson D.M.
        • Chen K.
        • Kutner M.H.
        • Wolfe R.A.
        The Dialysis Outcomes and Practice Patterns Study (DOPPS): design, data elements, and methodology.
        Am J Kidney Dis. 2004; 44: 7-15
        • Young E.W.
        • Goodkin D.A.
        • Mapes D.L.
        • et al.
        The Dialysis Outcomes and Practice Patterns Study: an international hemodialysis study.
        Kidney Int Suppl. 2000; 57: S74-S81
        • Pisoni R.L.
        • Bieber B.A.
        • Al Wakeel J.
        • et al.
        The Dialysis Outcomes and Practice Patterns Study phase 5 in the Gulf Cooperation Council countries: design and study methods.
        Saudi J Kidney Dis Transpl. 2016; 27: 1-11
        • Rayner H.C.
        • Greenwood R.
        • MacTier R.
        • et al.
        Estimated life expectancy of UK HD patients if clinical practice guidelines are met.
        Br J Renal Med. 2007; 12: 11-14
        • Robinson B.M.
        • Fuller D.S.
        • Dykstra D.M.
        • et al.
        The DOPPS practice monitor: rationale and methods for an initiative to monitor the new US bundled dialysis payment system.
        Am J Kidney Dis. 2011; 57: 822-831
        • Lopes A.A.
        • Albert J.M.
        • Young E.W.
        • et al.
        Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.
        Kidney Int. 2004; 66: 2047-2053
        • Klein J.
        • Moeschberger M.
        Survival Analysis: Techniques for Censored and Truncated Data.
        Springer, New York, NY1997: 416-418
        • Bonferroni C.E.
        Teoria statistica delle classi e calcolo delle probabilità.
        Pubblicazioni del R Istituto Superiore di Scienze Economiche e Commerciali di Firenze. 1936; 8: 3-62
        • Benjamini Y.
        • Hochberg Y.
        Controlling the false discovery rate: a practical and powerful approach to multiple testing.
        J R Stat Soc B. 1995; 57: 289-300
        • Raghunathan T.E.
        • Solenberger P.W.
        • Van Hoewyk J.
        IVEware: Imputation and Variance Estimation Software.
        Survey Methodology Program, Survey Research Center, Institute for Social Research, University of Michigan, Ann Arbor, MI2002
        • Foley R.N.
        • Gilbertson D.T.
        • Murray T.
        • Collins A.J.
        Long interdialytic interval and mortality among patients receiving hemodialysis.
        N Engl J Med. 2011; 365: 1099-1107
        • Bleyer A.J.
        • Russell G.B.
        • Satko S.G.
        Sudden and cardiac death rates in hemodialysis patients.
        Kidney Int. 1999; 55: 1553-1559
        • Zhang H.
        • Schaubel D.E.
        • Kalbfleisch J.D.
        • et al.
        Dialysis outcomes and analysis of practice patterns suggests the dialysis schedule affects day-of-week mortality.
        Kidney Int. 2012; 81: 1108-1115
        • Tentori F.
        • Blayney M.J.
        • Albert J.M.
        • et al.
        Mortality risk for dialysis patients with different levels of serum calcium, phosphorus, and PTH: the Dialysis Outcomes and Practice Patterns Study (DOPPS).
        Am J Kidney Dis. 2008; 52: 519-530
        • Robinson B.
        • Tong L.
        • Zhang J.
        • et al.
        Blood pressure levels and mortality risk among hemodialysis patients in the Dialysis Outcomes and Practice Patterns Study.
        Kidney Int. 2012; 82: 570-580
        • Jadoul M.
        • Thumma J.
        • Fuller D.S.
        • et al.
        Modifiable practices associated with sudden death among hemodialysis (HD) patients in the Dialysis Outcomes and Practice Patterns Study (DOPPS).
        Clin J Am Soc Nephrol. 2012; 7: 765-774
        • Wong M.M.Y.
        • McCullough K.P.
        • Bieber B.
        • et al.
        Interdialytic weight gain: trends, predictors, and associated outcomes in the international Dialysis Outcomes and Practice Patterns Study (DOPPS).
        Am J Kidney Dis. 2017; 69: 369-379
        • Karaboyas A.
        • Zee J.
        • Brunelli S.M.
        • et al.
        Dialysate potassium, serum potassium, mortality and arrhythmia events in hemodialysis: results from the DOPPS.
        Am J Kidney Dis. 2017; 69: 266-277
        • Ramirez S.P.B.
        • Kapke A.
        • Port F.
        • et al.
        Dialysis dose scaled to body surface area and size-adjusted, sex-specific patient mortalilty.
        Clin J Am Soc Nephrol. 2012; 7: 1977-1987
        • Obialo C.
        • Zager P.G.
        • Myers O.B.
        • Hunt W.C.
        Relationships of clinic size, geographic region, and race/ethnicity to the frequency of missed/shortened dialysis treatments.
        J Nephrol. 2014; 27: 425-430
        • Brar A.
        • Babakhani A.
        • Salifu M.O.
        • Jindal R.M.
        Evaluation of non-adherence in patients undergoing dialysis and kidney transplantation: United States transplantation practice patterns survey.
        Transplant Proc. 2014; 46: 1340-1346
        • Kutner N.G.
        • Zhang R.
        • McClellan W.M.
        • Cole S.A.
        Psychosocial predictors of non-compliance in haemodialysis and peritoneal dialysis patients.
        Nephrol Dial Transplant. 2002; 17: 93-99
        • Hedayati S.S.
        • Minhajuddin A.T.
        • Afshar M.
        • Toto R.D.
        • Trivedi M.H.
        • Rush A.J.
        Association between major depressive episodes in patients with chronic kidney disease and initiation of dialysis, hospitalization, or death.
        JAMA. 2010; 303: 1946-1953
        • Watnick S.
        • Kirwin P.
        • Mahnensmith R.
        • Concato J.
        The prevalence and treatment of depression among patients starting dialysis.
        Am J Kidney Dis. 2003; 41: 105-110
        • Weisbord S.D.
        • Mor M.K.
        • Sevick M.A.
        • et al.
        Associations of depressive symptoms and pain with dialysis adherence, health resource utilization, and mortality in patients receiving chronic hemodialysis.
        Clin J Am Soc Nephrol. 2014; 9: 1594-1602
        • DeOreo P.B.
        Hemodialysis patient-assessed functional health status predicts continued survival, hospitalization, and dialysis-attendance compliance.
        Am J Kidney Dis. 1997; 30: 204-212
        • Sabate E.
        Adherence to Long-term Therapies: Evidence for Action.
        World Health Organization, Geneva, Switzerland2003 (Accessed April 12, 2018)
        • Larsen J.
        • Stovring H.
        • Kragstrup J.
        • Hansen D.G.
        Can differences in medical drug compliance between European countries be explained by social factors: analyses based on data from the European Social Survey, round 2.
        BMC Public Health. 2009; 9: 145
        • Fissell R.B.
        • Karaboyas A.
        • Bieber B.A.
        • et al.
        Phosphate binder pill burden, patient-reported non-adherence, and mineral bone disorder markers: Findings from the DOPPS.
        Hemodial Int. 2016; 20: 38-49

      Linked Article

      • Missed Hemodialysis Treatments: A Modifiable But Unequal Burden in the World
        American Journal of Kidney DiseasesVol. 72Issue 5
        • Preview
          Of the approximately 2.5 million patients with end-stage renal disease (ESRD) on some form of renal replacement therapy worldwide, the vast majority (in most countries) receive in-center hemodialysis.1 In the United States, the frequency of missed hemodialysis treatments, barriers to better adherence with scheduled treatments, and consequences of missed dialysis treatments have been described.2-5 For instance, Chan et al3 investigated potential reasons behind missed dialysis sessions in the United States over the course of 5 years using data from a large dialysis organization.
        • Full-Text
        • PDF