Pain After Outpatient Surgical Procedures

A Survey of 330 000 Patients

Background: Approximately 8.1 million outpatient surgical procedures were performed in Germany in 2021. Little is known about the quality of postoperative pain treatment in the outpatient sector.

Methods: The AQS1 project comprises a combined survey of patients and staff in the framework of quality control for ambulatory surgery. The primary endpoint of this study was the prevalence of relevant incisional pain (≥ 4/10 on the numerical rating scale) up to postoperative day 3. Secondary endpoints included prognostic variables for pain and pain-associated outcomes, based on the AQS1 patient questionnaire. Moreover, mixed regression models were used to analyze potential prognostic variables and associations of pain with other outcomes (study registration number DRKS00028052).

Results: Data from 330 008 patients were evaluated (from 1 July 2001 to 31 December 2021). The overall prevalence of relevant incisional pain up to postoperative day 3 was 22.5%, with major differences between different types of procedure (3.2%–51.2%). Pain was most common after hemorrhoid surgery (51.2%) and the laparoscopic lysis of large and small bowel adhesions (45.4%). The main associations of relevant pain were with younger age (odds ratio [OR] 1.87, 95% confidence interval [1.82; 1.91]), early postoperative pain (1.34, [1.30; 1.39]), inadequate provision of analgesics (2.90, [2.71; 3.09]), and surgical wound infections (2.60, [2.43; 2.78]). Patients with pain reported lower overall satisfaction with the procedure and a longer inability to work.

Conclusion: These data have not been tested for representativeness. They can serve as a point of departure for the optimization of individualized perioperative pain therapy and for the planning of prospective studies.

Pain after surgical procedures and its prevention are a major concern in healthcare (1). According to Federal Health Reporting data, 8.1 million outpatient surgical procedures were performed in Germany in 2021 (2). While the quality of pain therapy in German hospitals is rather well researched thanks to the QUIPS project (QUIPS, Quality Improvement in Postoperative Pain Therapy [Qualitätsverbesserung in der postoperativen Schmerztherpie]) (3), data on the extent of postoperative pain after outpatient surgery are scarce (4, 5). Due to differences in the survey instruments used and the survey time periods evaluated, it is not possible to compare these findings with each other or with those of studies conducted in an inpatient setting. There is a complete lack of surveys comparing a relevant number of outpatient surgical procedures regarding postoperative pain or other outcomes.

For more than 20 years now, a number of facilities operated by statutory health insurance (SHI)-accredited physicians have been using the AQS1 system—a quality assurance instrument for outpatient surgical procedures on a voluntary basis (6), allowing participating institutions to carry out benchmarking and external quality control. The AQS1 system gathers demographic and procedure-related data; in addition, patients answer a postoperative questionnaire on the quality of treatment, including clinical outcome parameters such as postoperative pain (7).

The primary aim of this study was to analyze the prevalence of postoperative pain up to three days after surgery. Secondary endpoints included the evaluation of potential prognostic variables for this type of pain as well as associative analyses with regard to further relevant outcomes, such as inability to work.

Methods

Study design

This is a cross-sectional study based on anonymized data of the AQS1 registry (Quality Assurance in Outpatient Surgery). All procedures were performed in practices/outpatient surgery centers of SHI(statutory health insurance)-accredited physicians. The data were collected and digitalized for the purpose of quality assurance. The study was approved by the Ethics Committee of the Jena University Hospital (2 021–2 458 data) and is registered in the German Clinical Trials Register (DRKS00028052).

Data source

The AQS1 system comprises two parts: a documentation form and a patient questionnaire.

The documentation form is perioperatively filled by surgical/anesthesiological staff. Relevant process parameters are recorded, such as

• Risk factors
• Intraoperative complications
• The type of surgery performed (OPS code)
• Information about anesthesia and pain treatment.

Patients complete the patient questionnaire in paper form approximately two weeks after the outpatient surgical procedure. It contains questions on Patient Reported Outcome Measures (PROMs), such as pain intensity and patient satisfaction. A sample version of the AQS1 is available online as an eQuestionnaire (reprinted by courtesy of Medicaltex GmbH.)

Patient cohort

All adult patients recorded in the AQS1 registry between 1 July 2001 and 31 December 2021 were included. Only patients where a surgical procedure was performed and documented with a valid German operation and procedure code (OPS code “5-“) were included to ensure that they could later be linked to a specific procedure. Furthermore, only patients with valid information regarding the primary endpoint were included. In our current analysis, we aggregated surgical procedures into OPS groups that represent clinically similar procedures. In addition, complex procedures (multiple OPS codes during one operation) were included in these OPS groups. Detailed information on the methods used is provided in eTable 1 and eTable 2.

Variables

A list of all analyzed variables is provided in eTable 3. The primary endpoint was incisional pain up to postoperative day 3 (“Did you have any complaints due to incisional pain up to postoperative day 3?“, 0 = ”not at all“ up to 10 = ”unbearable“). The secondary endpoints (patient-reported) were:

• Adequate provision of analgesics in the recovery room and at home
• Complaints due to headache, neck pain and muscle pain after surgery
• The overall rating of the procedure
• Postoperative complications such as hemorrhage, infection, thrombosis
• Unplanned hospital admission.

Our analysis focused on the descriptive reporting of the relative prevalence of patients with at least moderate incisional pain up to postoperative day 3 (≥ 4/10 points on the numerical rating scale [NRS]: yes versus no) for the most common OPS groups (8, 9).

Statistical analyses

The median as well as the first (Q1) and third quartile (Q3) are reported for metric variables. For categorical variables, we report the absolute frequency and the relative frequency in percent. For all variables, the number of patients with valid data is stated.

For the primary endpoint, first the absolute and relative frequency of patients with at least moderate incisional pain up to postoperative day 3 (≥ 4/10 NRS: yes versus no) within each OPS group and the corresponding 95% confidence intervals are presented. In addition, the median as well as the first and second quartile of the point estimates of these percentages are reported across all OPS groups.

We used generalized linear mixed regression models in the associative analyses of prognostically relevant variables. A detailed overview is provided in the eBox. In summary, at least moderate incisional pain up to postoperative day 3 (yes versus no) was used as the dependent variable. In the base model, the following data were also modeled as independent variables:

• Age (≤ versus > OPS group-specific median split)
• Sex (male versus female)
• Classification of the American Society of Anesthesiologists (ASA)(I versus II versus III/IV),
• Urgency of surgery (plannable/elective versus urgent/emergency)
• Anesthesia administration (anesthesiologist versus surgeon)
• Incision-to-closure time (< versus ≥ OPS group-specific median split)
• Perioperative wound infiltration (yes versus no).

The extended model additionally included the following information as independent variables:

• Pain in the recovery room (yes versus no; surgeon-reported)
• Adequate provision of analgesics in the recovery room (yes versus no; patient-reported)
• Surgical wound infection (yes versus no; patient-reported)
• Postoperative thrombosis (yes versus no; patient-reported)
• Postoperative hematoma/postoperative bleeding (yes versus no; patient-reported).

The odds ratio (OR), including 95% confidence interval, and the p-value are reported for each variable.

Finally, we analyzed potential associations between at least moderate incisional pain up to postoperative day 3 (yes versus no; independent variable) and, as dependent variables, the following patient-reported information with regard to

• Postoperative inability to work (at least 1 day, yes versus no)
• Postoperative duration of inability to work (days)
• Emergency-related physician contact (yes versus no)
• Adequate provision of medication at home (yes versus no)
• Adequate care at home (yes versus no)
• Overall satisfaction with the procedure.

All analyses were performed using the R (version 4.1.0) and RStudio (version 1.4.1717) statistical software packages (10).

Results

A total of 798 120 cases were identified in the database. After exclusion of patients with missing information on age, surgical procedure (OPS code) and primary endpoint and after exclusion of OPS groups with less than 1000 cases, 330 008 patients were included in the analysis (Figure 1). A total of 34 OPS groups were analyzed (17 groups, 7 complex procedures within the groups and 10 specific individual procedures; Figure 2 and eTable 4).

Figure 1

Flowchart of the composition of the analysis sample

Figure 2

Descriptive presentation of the relative frequencies

Demographics and process parameters

The proportion of female patients was 65.7% (n = 215 661/328 281), the median age was 52 years (Q1–3: 41–62, n = 312 286) and the physical health status of the majority of patients was good (94.6% ASA I/II, n = 312 286).

The majority of the procedures included in the study were classified as elective (95.9%, n = 309 622/322 867) and only a small percentage as urgent (3.7%, within 7 days) or emergency (0.4%). In 11% of procedures (n = 35 254/321 159), anesthesia was administered by a surgeon, in the remainder of cases by an anesthesiologist (eTable 5). Overall, 71.7% of procedures were performed under general anesthesia (n = 236 530/330 008), 19.7% under regional/local anesthesia only and 5.7% under a combination. The median incision-to-closure time was 25 minutes (Q1–3: 14–40 minutes; n = 312 447).

Primary endpoint

In the total sample, the prevalence of at least moderate incisional pain (≥ 4 NRS) up to postoperative day 3 was 22.5% (n = 74 232/330 008). The prevalence rates varied significantly between the OPS groups analyzed (median: 22.2%, Q1–3: 17.2–34.1%, Figure 2). The highest prevalence rates were found for the surgical treatment of hemorrhoids (51.2%), laparoscopic lysis of adhesions on the small and large intestine (45.4%) and surgical procedures on the metatarsals and phalanges of the foot (44.9%). The lowest ranks were occupied by extracapsular cataract extraction (ECCE, 3.2%), turbinate surgery (submucosal resection, 12.7%) and abrasion of the uterus (curettage) (14.2%).

Secondary endpoints

Further complications

The rate of intraoperatively complications documented in the medical part was low (Table 1). Complications in the recovery room were also uncommon—apart from incisional pain with 11.5%—(circulatory problems, 1.5%; nausea,4%; vomiting, 0.8%; respiratory problems, 0.6%; allergic reactions, 0.5%; urinary retention, 0.5%; see eTable 6).

Table 1

Complications and patient-reported information during follow-up

Likewise, marked differences between the OPS groups with regard to further patient-reported postoperative complications, such as surgical wound infections (median: 2.7%, Q1–3: 1.5–3.5%; minimum: 0.7%, maximum: 7.8%, eFigure 1), thrombosis (median: 0.4%, Q1–3: 0.3–1.0%; minimum: 0.1%, maximum: 1.6%; eFigure 2) and hematoma/postoperative bleeding (median: 29.9%, Q1–3: 18.9–40.5%; minimum: 6.7%, maximum: 65.7%; eFigure 3), were only found for hematoma/postoperative bleeding.

Most patients were positive in their assessment of the quality of the outcome. This is reflected in the good to very good overall assessment of the procedure in 98.2% of cases. A detailed overview of the results is provided in Table 1.

Variables relevant to prognosis

The results of the basic model (n = 285 478) are shown in Table 2. Younger age, longer incision-to-closure time, female sex, higher ASA classification, and increased urgency of the procedure were associated with an increased risk for at least moderate incisional pain up to postoperative day 3. Anesthesia administration by surgeons and perioperative administration of local anesthetics (at least one administration pre-, intra- or postoperatively) were associated with a reduced risk. Most of the effect sizes of the variables analyzed are small.

Table 2

Results of the regression models

In an extended model (Table 2, n = 193 828), pain in the recovery room (surgeon-reported) and subsequent complications, such as surgical wound infection, thrombosis and hematoma/postoperative bleeding, were also associated with an increased probability of occurrence. If patients stated that they had been given analgesics in adequate amounts, this was associated with a reduced likelihood of pain. The effects of the newly included factors in particular were at a medium level. In the extended model, the urgency of surgery and the perioperative administration of local anesthetics showed no association with the outcome. Further details are provided in eTable 7.

As a limitation, it should be noted that the sample size was considerably smaller for the extended model compared to the base model (58.7% versus 86.5% of the total sample, n = 285 478).

Associations with further relevant outcome parameters

In the regression models, at least moderate incisional pain up to postoperative day 3 was associated with an increased rate of inability to work (OR: 2.21, [2.07; 2.34], p<0.001), longer duration of inability to work (12.5 d, [9.7; 15.4 d] versus 14.7 d, [11.8; 17,5], p<0.001) and an increased rate of emergency-related physician contacts (OR: 2.66, [2.49; 2.83], p<0.001). In addition, the occurrence of pain was associated with poorer outcomes for feeling adequately cared for at home (OR: 0.37, [0.36; 0.39], p<0.001) or being adequately supplied with medications at home (OR: 0.35, [0.34; 0.37], p<0.001), as well as lower overall satisfaction with the procedure (OR: 0.47, [0.47; 0.48], p<0.001). In summary, the effect sizes were in the low to moderate range (details provided in eTable 8).

Discussion

In this study, we investigated for the first time a broad spectrum of outpatient surgical procedures with regard to incisional pain by analyzing a very large number of PROM data sets. Approximately one quarter of patients who underwent outpatient surgery reported moderate to severe pain. Considerable differences were found between the surgical procedures included in our analysis. In ten groups of surgical procedures, including very frequently performed procedures such as plasty of the capsular ligaments in the shoulder joint and hernia surgery, one in three patients reported moderate to severe pain.

Our study thus confirms the wide range in the frequency of pain after outpatient surgical procedures which in the literature spans from 24% to 82% (11, 12, 13, 14). Gerbershagen et al. (2013) already showed in a study on 179 inpatient surgical procedures that even supposedly minor procedures can be associated with severe pain (15). They found that, for example, tonsillectomy and appendectomy as well as shoulder operations were much more painful than some major procedures. However, the authors emphasize that the perceived pain intensity does not only depend on the procedure performed but also on adequate pain treatment. The patient risk factors younger age and female sex have already been identified in multiple studies (see sensitivity analysis in eTable 9) (16, 17, 18, 19, 20, 21, 22).

The results of our analysis should prompt efforts to optimize perioperative pain treatment in patients who underwent outpatient surgical procedures identified as being particularly painful. Our analysis showed an association between the perioperative administration of local anesthetics and a reduced risk of pain. Regional analgesia techniques are superior to systemic pain therapy alone in patients undergoing shoulder surgery (23), while additional infiltration of local anesthetics should be performed in patients undergoing hemorrhoidectomy (23); for hernia repair, systemic analgesia should be combined with local or regional analgesia techniques (24). These publications also emphasize the analgesic component of perioperative dexamethasone administration in all surgical procedures discussed here. Furthermore, there is good evidence supporting the acute analgesic effect of transcutaneous electrical nerve stimulation (TENS) after various surgical procedures. While this technique is still not commonly used, it is ideally suited for applications in the outpatient setting. It is in line with the principle of “patient-controlled” analgesia and has hardly any contraindications (23, 25). TENS therapy is especially useful in orthopedic and trauma surgery and abdominal surgery.

While in an inpatient setting pain treatment can be intensified and patients can be seen by medical staff at all times, the situation is more complex in the home environment. Thus, preoperative risk assessment is of particular importance in the outpatient setting as it helps to plan the method of pain treatment in advance and to enable patients to manage complaints they may be experiencing. Moderate to severe pain after surgery is associated with a poorer quality of life and delayed postoperative recovery. This is reflected in our results in increased and prolonged inability to work and more frequent emergency contacts and supported by the literature (12, 26, 27). Moreover, the association between pain-related variables and overall satisfaction (28) and process satisfaction is evident in the available data, providing a further argument for paying more attention to pain treatment after outpatient surgical procedures.

From a health policy perspective, it is important to emphasize that the AQS1 project—similar to QUIPS—is one of the few external, independent quality assurance instruments in Germany that focuses on the patient perspective and is carried out on a voluntary basis by healthcare providers as part of their daily routine. Apart from enabling institutional benchmarking, the large amount of data allows the analysis of numerous healthcare science and clinical research questions. With this study, we intend to provide a starting point for such analyses.

Strengths and limitations

Our analysis is the by far largest evaluation of patient-reported outcome measures after outpatient surgical procedures. Previous studies analyzed much smaller numbers of cases and are usually limited to selected surgical procedures (12, 26, 27). Further major strengths of our study include the evaluation of multiple relevant complications, such as pain, thrombosis, surgical wound infection, and postoperative bleeding across a broad spectrum of surgical procedures. By comparatively analyzing numerous surgical procedures based on data collected with one and the same instrument, we were able to identify for the first time those procedures that are associated with an above-average risk of intense postoperative pain.

The response rate to the questionnaires was about 40%. Information about patients who did not return the questionnaire is not available. Comparative studies have shown that non-responders tend to have a poorer postoperative outcome (29, 30). For this reason, we assume that satisfied patients with good outcomes were overrepresented in our study. The data made available to us did not allow to analyze center-specific effects and changes in outcome over time during the observation period. We assume that there are considerable regional differences in postoperative outcomes. No claim can be made as to the representativeness of the data. Like any retrospective evaluation of registry data, our analysis does not allow to draw any conclusions about causal relationships.

Conclusion

In our study, outpatient surgical procedures were associated with low complication rates and a generally high level of patient satisfaction. Still, about one quarter of patients who underwent outpatient surgery reported moderate to severe postoperative incisional pain, with the numbers varying significantly between different procedures.

Financial support

The Jena University Hospital received financial support from the German Federal Association of Ambulatory Surgery (Bundesverband Ambulantes Operieren e.V., BAO) as part of the joint research project “Pain Treatment After Outpatient Surgical Procedures: Analysis of the AQS1 Survey“.

Conflict of interest statememt
KB is the managing partner of Medicaltex GmbH.

WM received consulting fees from Tafalgie, Grünenthal, Ethypharm, and Mundipharma and lecture fees from Grünenthal, Munipharma, Spectrum Therapeutics, and Kyowa. He is a member of the Executive Committee of the German Pain Society (Deutsche Schmerzgesellschaft).

AN is Vice President of the German Federal Association of Outpatient Surgery (BAO, Bundesverbandes für Ambulantes Operieren e. V.).

The remaining authors declare that no conflict of interest exists.

Manuscript received on 23 June 2023; revised version accepted on 25 October 2023.

Translated from the original German by Ralf Thoene, M.D.

Corresponding author
Apl. Prof. Dr. med. Winfried Meißner
Klinik für Anästhesiologie und Intensivmedizin
Universitätsklinikum Jena
Am Klinikum 1, 07747 Jena, Germany

Cite this as:
Baumbach P, Dreiling J, Arnold C, Weinmann C, Komann M, Bäcker K, Neumann A, Karst J, Meißner W: Pain after outpatient surgical procedures—a survey of 330 000 patients. Dtsch Arztebl Int 2024; 121: 71–8. DOI: 10.3238/arztebl.m2023.0235

References

1. Meissner W, Coluzzi F, Fletcher D, et al.: Improving the management of post-operative acute pain: priorities for change. Curr Med Res Opin 2015; 31: 2131–43

2. Gesundheitsberichterstattung des Bundes: Art der ärztlichen Leistung für Versicherte der Gesetzlichen Krankenversicherung, Abrechnungsfälle/Leistungsfälle: Ambulantes Operieren. https://www.gbe-bund.de/gbe/pkg_isgbe5.prc_menu_olap?p_uid=gast&p_aid=99037419&p_sprache=D&p_help=2&p_indnr=263&p_indsp=6146&p_ityp=H&p_fid= (last accessed on 7 June 2023).

3. Meißner W, Komann M, Erlenwein J, Stamer U, Scherag A: The quality of postoperative pain therapy in German hospitals—the effect of structural and procedural variables. Dtsch Arztebl Int 2017; 114: 161–7

4. Lux E, Stamer U, Meissner W, Wiebalck A: Postoperative Schmerztherapie nach ambulanten Operationen. Der Schmerz 2011; 2: 191–8

5. Schwarze C, Zenz D, Orlowski O, et al.: Erhebung von Schmerzen nach ambulanten Operationen. Der Schmerz 2016; 2: 141–51

6. Rüggeberg J, Brökelmann J: Qualitätssicherung AQS1 hat sich für das Ambulante Operieren bewährt. Ambulant operieren 2004; 4: 151–6.

7. Bäcker K: 3.3 AQS1–Patientensicherheitsfragebogen: Qualitätsdokumentation, Langzeitergebnisse und Versorgungsforschung. Manual Ambulantes Operieren: de Gruyter 2016; p. 150–61

8. Gerbershagen HJ, Rothaug J, Kalkman CJ, Meissner W: Determination of moderate-to-severe postoperative pain on the numeric rating scale: a cut-off point analysis applying four different methods. Br J Anaesth 2011; 107: 619–26

9. Mendoza TR, Chen C, Brugger A, et al.: Lessons learned from a multiple-dose post-operative analgesic trial. Pain 2004; 109: 103–9

10. R Core Team: R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing 2021.

11. Apfelbaum JL, Chen C, Mehta SS, Gan TJ: Postoperative pain experience: results from a national survey suggest postoperative pain continues to be undermanaged. Anesth Analg 2003; 97: 534–40

12. McGrath B, Elgendy H, Chung F, Kamming D, Curti B, King S: Thirty percent of patients have moderate to severe pain 24 hr after ambulatory surgery: a survey of 5,703 patients. Can J Anaesth 2004; 51: 886–91

13. Pavlin DJ, Chen C, Penaloza D, Polissar NL, Buckley FP: Pain as a factor complicating recovery and discharge after ambulatory surgery. Anesth Analg 2002; 95: 627–34

14. Segerdahl M, Warrén-Stomberg M, Rawal N, Brattwall M, Jakobsson J: Clinical practice and routines for day surgery in Sweden: results from a nation-wide survey. Acta Anaesthesiol Scand 2008; 52: 117–24

15. Gerbershagen HJ, Aduckathil S, van Wijck AJ, Peelen LM, Kalkman CJ, Meissner W: Pain intensity on the first day after surgery: a prospective cohort study comparing 179 surgical procedures. Anesthesiology 2013; 118: 934–44

16. Tighe PJ, Le-Wendling LT, Patel A, Zou B, Fillingim RB: Clinically derived early postoperative pain trajectories differ by age, sex, and type of surgery. Pain 2015; 156: 609

17. Tighe PJ, Riley III JL, Fillingim RB: Sex differences in the incidence of severe pain events following surgery: a review of 333,000 pain scores. Pain medicine 2014; 15: 1390–404

18. Bartley EJ, Fillingim RB: Sex differences in pain: a brief review of clinical and experimental findings. Br J Anaesth 2013; 111: 52–8

19. Fillingim RB, King CD, Ribeiro-Dasilva MC, Rahim-Williams B, Riley III JL: Sex, gender, and pain: a review of recent clinical and experimental findings. J Pain 2009; 10: 447–85

20. Ruau D, Liu LY, Clark JD, Angst MS, Butte AJ: Sex differences in reported pain across 11,000 patients captured in electronic medical records. J Pain 2012; 13: 228–34

21. Weinmann C, Komann M, Meissner W: Tough cookies: the older the patients, the more pain tolerating. Eur J Anaesthesiol 2017; 34: 215.

22. Zheng H, Schnabel A, Yahiaoui-Doktor M, et al.: Age and preoperative pain are major confounders for sex differences in postoperative pain outcome: a prospective database analysis. PLoS One 2017; 12: e0178659

23. Deutsche Gesellschaft für Anästhesiologie und Intensivmedizin e.V. (DGAI): Behandlung akuter perioperativer und posttraumatischer Schmerzen, Version 4.1 vom 25.11.2022. https://www.register.awmf.org/de/leitlinien/detail/001-025 (last accessed on 7 June 2023).

24. Joshi GP, Rawal N, Kehlet H, et al.: Evidence-based management of postoperative pain in adults undergoing open inguinal hernia surgery. Br J Surg 2012; 99: 168–85

25. Johnson MI, Paley CA, Howe TE, Sluka KA: Transcutaneous electrical nerve stimulation for acute pain. Cochrane Database Syst Rev 2015; 2015: CD006142

26. Beauregard L, Pomp A, Choinière M: Severity and impact of pain after day-surgery. Can J Anaesth 1998; 45: 304–11

27. Watt-Watson J, Chung F, Chan VW, McGillion M: Pain management following discharge after ambulatory same-day surgery. J Nurs Manag 2004; 12: 153–61

28. Schwenkglenks M, Gerbershagen HJ, Taylor RS, et al.: Correlates of satisfaction with pain treatment in the acute postoperative period: results from the international PAIN OUT registry. Pain 2014; 155: 1401–11

29. Hutchings A, Frie KG, Neuburger J, van der Meulen J, Black N: Late response to patient-reported outcome questionnaires after surgery was associated with worse outcome. J Clin Epidemiol 2013; 66: 218–25

30. Imam MA, Barke S, Stafford GH, Parkin D, Field RE: Loss to follow-up after total hip replacement: a source of bias in patient reported outcome measures and registry datasets? Hip Int 2014; 24: 465–72

e1. Nakagawa S, Johnson PCD, Schielzeth H: The coefficient of determination R 2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded. J R Soc Interface 2017; 14: 20170213