Article

Gestalt for shock and mortality in the emergency department: A prospective study

a b s t r a c t

Objective: The diagnosis of shock in patients presenting to the emergency department (ED) is often challenging. We aimed to compare the accuracy of experienced emergency physician gestalt against Li’s pragmatic shock (LiPS) tool for predicting the likelihood of shock in the emergency department, using 30-day mortality as an ob- jective standard.

Method: In a prospective observational study conducted in an urban, academic ED in Hong Kong, adult patients aged 18 years or older admitted to the resuscitation room or high dependency unit were recruited. Eligible pa- tients had a standard ED workup for shock. The emergency physician treating the patient was asked whether he or she considered shock to be probable, and this was compared with LiPS. The proxy ‘gold’ or reference stan- dard was 30-day mortality. The area under the receiver operating curve (AUROC) was used to predict prognosis. The primary outcome measure was 30-day mortality.

Results: A total of 220 patients fulfilled the inclusion criteria and were included in the analysis. The AUROC for LiPS (0.722; sensitivity = 0.733, specificity = 0.711, P b 0.0001) was greater than emergency physician gestalt (0.620, sensitivity = 0.467, specificity = 0.774, P = 0.0137) for diagnosing shock using 30-day mortality as a proxy (dif- ference P = 0.0229). LiPS shock patients were 6.750 times (95%CI = 2.834-16.076, P b 0.0001) more likely to die within 30-days compared with non-shock patients. Patients diagnosed by emergency physicians were 2.991 times (95%CI = 1.353-6.615, P = 0.007) more likely to die compared with the same reference.

Conclusions: LiPS has a higher diagnostic accuracy than emergency physician gestalt for shock when compared against an outcome of 30-day mortality.

(C) 2017

Introduction

Shock is often difficult to diagnose in the emergency department (ED) and delays in recognition and treatment adversely affect patient survival [1,2]. At the bedside, doctors identify shock using clinical histo- ry, health status, vital signs, examination, and easily available investiga- tions such as urine output and blood lactate [3,4]. Appling a clinical diagnostic algorithm using these components which are generated based on the physician expertise is a way to improve diagnostic accura- cy [5,6]. Recently there has been increased interest in evaluating the role of clinician experience, otherwise known as gestalt, in the assessment of

? Disclosure: there are no conflicts of interest.

* Corresponding author at: Emergency Medicine Academic Unit, Cardiff University, University Hospital of Wales, Heath Park, Cardiff, UK.

E-mail addresses: [email protected] (C.A. Graham), [email protected] (T.H. Rainer).

disease [7] or prediction of the treatment failure [8], and some studies have concluded that emergency physician gestalt is not sufficiently ac- curate [9,10]. Thus, there is a need to investigate emergency physician gestalt for diagnosing shock in the ED phase of care [11].

Accepting that patients present along a spectrum of clinical severity and that shock lies on this spectrum, we have recently validated a prag- matic tool – Li’s a priori Pragmatic Shock (LiPS) strategy – for identifying possible and probable shock in the emergency setting. This strategy consists of the following components, systolic blood pressure (SBP), mean arterial pressure (MAP), Clinical impression of tissue perfusion, blood lactate, pH and Base deficit. Furthermore, when shock presents, the patient is classified according to peripheral skin temperature (see Table 1) [3]. When applying this novel protocol in the clinical setting, we need to demonstrate that it is at least as good as clinical gestalt, if not better. A proxy for the difference between clinical gestalt and LiPS for determining shock in the ED is differences in 30-day mortality. We considered that the difference between clinical gestalt and LiPS was

https://doi.org/10.1016/j.ajem.2017.11.007 0735-6757/(C) 2017

Table 1

Clinical manifestation and laboratory results of patients grouped by LiPS (N = 220).

Variable

N

No shock

Possible shock

Cold peripheries shock

Warm peripheries shock

Hot peripheries shock

P value

N = 37

N = 106

N = 32

N = 40

N = 5

First observations

Age (y)

220

65.89 +- 18.56

70.36 +- 16.26

72.06 +- 12.76

67.33 +- 14.50

72.80 +- 14.22

0.408

Temperature (?C)

220

37.01 +- 0.73

36.68 +- 0.90

36.13 +- 1.12?,?,?

37.15 +- 1.16

38.88 +- 1.13?,?,?,?

b 0.001

SBP (mmHg)

220

144.41 +- 32.99

143.16 +- 29.93

109.59 +- 32.87?,?

125.13 +- 40.32?

93.60 +- 20.99?,?

b 0.001

DBP (mmHg)

220

74.68 +- 17.21

76.94 +- 20.15

61.93 +- 18.14?

69.10 +- 21.38

54.00 +- 6.82

b 0.001

(mmHg)

220

97.92 +- 20.76

99.02 +- 21.01

77.82 +- 21.85?,?

87.77 +- 25.73

67.20 +- 10.40?,?

b 0.001

Shock Index

220

0.68 +- 0.32

0.65 +- 0.29

0.88 +- 0.32?

0.81 +- 0.36

1.21 +- 0.28?,?

b 0.001

Laboratory results

pH

211?

7.42 +- 0.04

7.40 +- 0.07

7.36 +- 0.11?

7.35 +- 0.11?,?

7.44 +- 0.07

0.001

BE

211?

1.16 +- 2.70

0.54 +- 3.89

-4.68 +- 5.08?,?

-5.84 +- 5.97?,?

-2.66 +- 2.79

b 0.001

WBC (x10^9/L)

220

9.28 +- 5.62

9.63 +- 4.07

10.84 +- 5.21

13.77 +- 8.07?,?

10.34 +- 5.95

0.001

RBC (x10^9/L)

220

3.91 +- 0.90

4.15 +- 0.86

3.61 +- 1.01?

3.83 +- 1.19

4.01 +- 0.50

0.050

Hemoglobin (g/dl)

220

11.50 +- 2.90

12.38 +- 2.72

9.99 +- 2.53?

11.14 +- 3.10

11.38 +- 2.15

0.001

Urea (mmol/L)

219?

8.07 +- 6.00

8.14 +- 6.36

17.85 +- 14.32?,?

13.20 +- 9.51?

9.38 +- 4.27

b 0.001

Lactate (mmol/L)

214?

1.18 +- 0.20

2.17 +- 0.65

3.84 +- 3.13?,?

4.43 +- 3.27?,?

4.02 +- 2.18?

b 0.001

Adverse outcome

30-day mortality

220

0 (0)

8 (7.55)

15 (46.88)?,?,?

6 (15.00)?

1 (20.00)

b 0.001

n (%)

30-day ICU admission

220

4 (10.81)

14 (13.21)

6 (18.75)

8 (20.00)

4 (80.00)?,?,?,?

0.0012

n (%)

In the overall group analysis, Kruskall-Wallis was used to evaluate any significant differences between groups.

? Missing data of three patients from cold peripheries shock group and three patients from warm peripheries shock.

? Missing data of two patients from possible shock group, five patients from cold peripheries shock group and two patients from warm peripheries shock group.

? Comparison between “no shock” and “cold peripheries shock” group, P b 0.05.

? Comparison between “no shock” and “warm peripheries shock” group, P b 0.05.

? Comparison between “no shock” and “hot peripheries shock” group, P b 0.05.

? Comparison between “possible shock” and “cold peripheries shock” group, P b 0.05.

? Comparison between “possible shock” and “warm peripheries shock” group, P b 0.05.

  • Comparison between “possible shock” and “hot peripheries shock” group, P b 0.05.
  • Comparison between “cold shock” and “warm peripheries shock” group, P b 0.05.

? Comparison between “cold shock” and “hot peripheries shock” group, P b 0.05.

? Comparison between “warm peripheries shock” and “hot peripheries shock” group, P b 0.05.

significant if the area under the receiver operator curve or the odds ratio for predicting 30-day mortality was statistically significantly different.

Thus, we aimed to determine the accuracy of experienced emergen- cy physician gestalt for detecting and predicting likelihood of shock in the ED compared with LiPS using 30-day mortality as the gold or refer- ence standard.

Materials and methods

Study design and eligibility

The Clinical Research Ethics commission of The Chinese University of Hong Kong approved a prospective, single-centre study in the ED of the Prince of Wales Hospital, Hong Kong, SAR, China. Written consent was obtained from each patient or their relatives. Patient enrolment oc- curred between July 2012 and August 2014, 9 am to 4 pm on each week- day. Adult patients who were at least 18 years old and presented to two resuscitation rooms or three high dependency units (HDUs) were re- cruited. The HDU is a unit of three separate cubicles with continuous cardiac monitoring located in ED. It can provide intensive and emergen- cy medical care for patients. In a five-point triage scale where category 1 is critical (immediate assessment), category 2 is emergency (medical assessment within 15 min), category 3 is urgent (medical assessment within 30 min), category 4 is semi-urgent, and category 5 is non-urgent, the HDU will accept category 2 or 3 depending on available beds and staff. The exclusion criteria were: b 18 years old, pregnant, and breastfeeding.

Data collection and definitions

Two cohorts of data were collected for the present study, 111 pa- tients from the original publication [3] and a further 109 patients from a subsequent unpublished dataset.

When assessing patients, emergency physicians gave their opinion, based on all available data in the emergency department, whether they thought the patient had shock, and if so, the type of shock. Accord- ing to LiPS, patients were first classified into one of three groups: no, possible or probable shock. In the probable shock group, patients were further divided according to peripheral (hand) temperature – cold, warm or hot peripheries [3,12].

Questions

In the emergency department, physician doctors were asked: Did you think that this patient is in shock? If yes, what kind of shock do you think that the patient has?

Outcome measures

The primary outcome was 30-day all-cause mortality.

Statistical analysis

All analyses were conducted using Medcalc version 15.8. Quan- titative parameters are presented as mean +- standard deviation (SD). Comparisons of quantitative parameters were analyzed by one-way analysis of variance (ANOVA) or Kruskal-Wallis tests. Bonferroni tests between groups were conducted if the P value of the ANOVA tests were b 0.05. Qualitative parameters were ana- lyzed by Chi-square test as appropriate. Area under the receiver operating characteristic (AUROC) curve analyses was performed to predict prognosis. Odds ratios (ORs) with 95% CIs were calcu- lated. A P-value of b 0.05 was considered to be statistical significant.

Results

Study participants

During two data collection periods – July 2012 to January 2013 and December 2013 to August 2014 – a total of 220 patients (mean age 69.36 +- 15.88 years; male 62.73%) fulfilled the inclusion criteria and were evaluated. Of these 220 patients, 37 (16.82%) patients were classi- fied as ‘No Shock’,106 (48.18%) patients as ‘Possible Shock’,32 (14.55%) patients as ‘Cold Peripheries Shock’, 40 (18.18%) patients as ‘Warm Pe- ripheries Shock’ and five (2.27%) patients as ‘Hot Peripheries Shock’.

Comparison of clinical symptoms, laboratory results and outcome

Table 1 shows the comparison of clinical variables consisted in the LiPS definition among the five groups. There were significant differences between groups. Cold peripheries shock and hot peripheries shock group had lower SBP and MAP, whilst cold peripheries shock and warm peripheries shock patients had lower pH and BE. The probable shock group patients had higher lactate. Among the five groups, the cold peripheries shock patients had the highest 30-day mortality (46.88%, P b 0.01), whilst the hot peripheries shock cases were more likely to be sent to ICCU (80%, P = 0.0012).

Accuracy analysis

Table 2 and Fig. 1 shows the accuracy of emergency physicians’ ge- stalt for shock compared with LiPS. The AUROCs of emergency physi- cians and LiPS were 0.620 (95%CI = 0.552-0.685, P = 0.0137) and 0.722 (95%CI = 0.658-0.780, P <= 0.0001) respectively (difference

0.102 (P = 0.0229)).

When the LiPS ‘possible shock group’ is excluded (N = 106), Table 3 shows the agreement in shock assessment between emergency physi- cians gestalt and LiPS (N = 114). In 74/114 cases, there was agreement between doctors’ gestalt and LiPS. Of 37 patients who were classified as ‘probable shock’ by LiPS but ‘no shock’ by emergency physicians, 32 cases had either a metabolic acidosis (N = 22) or a high lactate level (N = 17) or both metabolic acidosis and a high lactate level (N = 5). 32 patients had a systolic pressure N 90 mmHg and 34 patients had a mean arterial blood pressure N 65 mmHg. Eight of 37 (21.6%) patients in this subgroup died within 30 days.

Table 4 presents the odds ratios for 30-day mortality. Emergency physicians (OR = 2.991, 95%CI = 1.353-6.615, P = 0.007) were less likely to identify probable shock than LiPS (30-day mortality OR = 6.750, 95%CI = 2.834-16.076, P b 0.0001). Especially in cold peripheries shock patients diagnosed by LiPS, 30-day mortality was 10 times higher than the non-cold shock patients (95%CI = 4.255-24.339, P b 0.0001).

Discussion

This study has extended research from our previous study. To our knowledge, we were the first team to use objective ED-based clinical variables to develop an a priori definition of shock i.e. LiPS. LiPS not only included a classification of shock, but also the severity of shock.

Fig. 1. Comparison of predictive accuracy against 30-day mortality.

The clinical gestalt for shock in the ED was a holistic clinical impression of patients made by emergency physician, which was partly subjective and susceptible to interference [13]. The accuracy and variability of a doctors’ judgment which comprises a complex and immeasurable num- ber of variables is difficult to determine [14]. This study aimed to com- pare the accuracy of emergency physicians’ gestalt with LiPS for diagnosing probably shock, whilst using 30-day mortality as an objec- tive proxy reference.

We first compared the prediction accuracy difference by AUROC tests of 220 patients. Emergency physicians could make their judgments based on available information in the emergency department. Applying 30-day mortality as a reference standard, emergency physicians identi- fied b 50% patients (14/30) with probable shock. On the other hand, emergency physicians were more accurate at identifying patients with- out shock (147/190), with a specificity which was slightly higher than LiPS. However, overall, LiPS had a higher sensitivity than clinical gestalt. Emergency physicians are under work and time pressure to make a diagnostic decision and to correctly manage critically ill patients, and this is often based on a limited amount of material and on clinical expe- rience [15]. Doctor may misunderstand the Severity of disease [16], and may take no account of laboratory results [17], which have been proven to predict adverse outcome [18-20]. Further evidence to support this may be found in the 37 patients classed as probable shock by LiPS but who were classified as ‘no shock’ by emergency physicians, and most of whom had a systolic pressure N 90 mmHg or mean artery blood pres- sure N 65 mmHg. The relatively normal blood pressure may be interpreted as a stable and relatively low risk patient. An increased probability of shock could have been noted not only by a high lactate

Table 2

Comparison of predictive accuracy against 30-day mortality (N = 220).

Shock

30-day mortality

AUROC

Sensitivity

Specificity

P value

No Yes

(95%CI)

Dr’s gestalt

No

147 16

0.620

0.467

0.774

0.0137

Yes

43 14

(0.552-0.685)

(0.283-0.657)

(0.708-0.831)

LiPSa

No/possible

135 8

0.722

0.733

0.711

b0.0001

Yes

55 22

(0.658-0.780)

(0.541-0.877)

(0.640-0.774)

Difference

0.102

0.0229

(0.014-0.189)

a Compare with LiPS probable shock vs no/possible shock.

Table 3

Agreement in shock assessment between Dr.’s gestalt and LiPS (N = 114).

Dr’s Gestalt shock

No Yes

have not compared LiPS against other potential early risk tools such as NEWS or qSOFA.

Despite these limitations, our study shows some interesting find- ings. The LiPS was more accurate in diagnosing shock than experi- enced ED medical gestalt when compared against an outcome of

LiPS shock No 34

Yes 37a

a 5 patients had a SBP b 90 mmHg; 3 patients had a MAP b 65 mmHg.

3 30-day mortality. Further, patients fulfilling the LiPS definition to

40 be diagnosed with shock were 6 times more likely to die in 30 days, especially in those patients who were diagnosed with cold peripheries shock.

level but alternatively by the degree of metabolic acidosis. Presently there is a high focus on lactate levels in critically ill ED patients but less focus on metabolic acidosis, despite the fact that acidosis probably reflects shock and tissue hypoxia more accurately than lactate, since el- evated lactate may be related to drugs or toxins which is independent to shock or hypoxia [21,22]. The majority of these patients were in meta- bolic acidosis not hyperlactacidemia. The mortality in this group is sig- nificant as eight (21.6%) patients died within 30 days, and were not admitted to ICCU. These patients may have been overlooked because laboratory results were either not available at an early stage, or because their significance was not appreciated. Our data suggests that metabolic acidosis and decreased bicarbonate levels may alert clinicians to the possibility of shock in the presence of normal or only moderately elevat- ed lactate, and should be incorporated in future guidelines and assessments.

We have developed a pragmatic tool for shock assessment

whereby cold peripheries usually suggest vasoconstrictive shock and warm peripheries suggest Vasodilatory shock. Although this as- sociation is not absolute, in our previous study [3] we found that this categorization did equate to very different mortality outcomes. There are undoubtedly exceptions to this rule whereby cold hands equate to vasoconstriction and warm hands equate to vasodilatation. However, until a reliable technique for assessing systemic vascular resistance such as USCOM is easily and readily available in ED, and incorporated into shock assessment, the rule applies in the majority of cases. In fact, we have confirmed this in another unpublished study (TR, personal communication).

The subgroup of cold peripheries shock patients classified using LiPS, most of whom were considered to be in late stages of shock, had an odds ratio of 10 for 30-day mortality. Thus, LiPS has the potential to identify both the presence and severity of shock.

Following the Diagnostic protocol, and using point of care tests, it should not take N 10 min to determine the probability of shock in the ED in a given patient. However, our study identified parameters, which may further improve the accuracy of shock and could be evaluat- ed in further refinements of LiPS. These include temperature, shock index, SBP-DBP difference, hemoglobin, white cell count and renal func- tion. In addition, non-invasive hemodynamic devices have the ability to measure systemic vascular resistance, cardiac output, preload, cardiac power and oxygen delivery, all of which may further improve shock di- agnosis [23,24].

A number of limitations in our study should be acknowledged. Though we extended our previous preliminary study with 220 patients, the current research was a single-centre study. The result of this study might be influenced by the working system of the emergency depart- ment, and the patient source composition. In our opinion, though the trend of advantage of applying the LiPS in emergency department was clear, it is still necessary to conduct a multi-centre study. Further, we

Table 4

Odds ratios for 30-day mortality (N = 220).

Variables (vs reference) OR

95% CI

P value

30-day mortality

Dr’s opinion shock (vs no shock)

2.991

1.353-6.615

0.007

LiPS probable shock (vs no/possible shock)

6.750

2.834-16.076

b 0.0001

LiPS cold shock (vs no cold shock)

10.176

4.255-24.339

b 0.0001

Conclusion

In our ED-based pragmatic study, LiPS was more accurate overall than clinical gestalt for determining shock using 30-day mortality as an objective proxy reference. The LiPS tool may be used in the ED for im- proving the diagnosis of shock and for research.

References

  1. Vincent JL, Rhodes A, Perel A, Martin GS, Della Rocca G, Vallet B, et al. Clinical re- view: update on hemodynamic monitoring–a consensus of 16. Crit Care 2011; 15(4):229.
  2. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman CS, et al. Developing a new definition and assessing new clinical criteria for septic shock: for the third international consensus definitions for sepsis and septic shock (sep- sis-3). JAMA 2016;315(8):775-87.
  3. Li YL, Chan CP, Sin KK, Chan SS, Lin PY, Chen XH, et al. Validating a pragmatic defi- nition of shock in adult patients presenting to the ED. Am J Emerg Med 2014; 32(11):1345-50.
  4. El Ayadi AM, Nathan HL, Seed PT, Butrick EA, Hezelgrave NL, Shennan AH, et al. Vital sign prediction of adverse maternal outcomes in women with hypovolemic shock: the role of shock index. PLoS One 2016;11(2):e0148729.
  5. Ramanayake RP, Ranasingha S, Lakmini S. Management of emergencies in gener- al practice: role of general practitioners. J Family Med Prim Care 2014;3(4): 305-8.
  6. Simmons J, Ventetuolo CE. Cardiopulmonary monitoring of shock. Curr Opin Crit Care 2017;23(3):223-31.
  7. Balamuth F, Alpern ER, Grundmeier RW, Chilutti M, Weiss SL, Fitzgerald JC, et al. Comparison of two sepsis recognition methods in a pediatric emergency depart- ment. Acad Emerg Med 2015;22(11):1298-306.
  8. Rippey JC, Carr PJ, Cooke M, Higgins N, Rickard CM. Predicting and preventing pe- ripheral intravenous cannula insertion failure in the emergency department: clini- cian ‘gestalt’ wins again. Emerg Med Australas 2016;28(6):658-65.
  9. Body R, Cook G, Burrows G, Carley S, Lewis PS. Can emergency physicians ‘rule in’ and ‘rule out’ acute myocardial infarction with clinical judgement? Emerg Med J 2014;31(11):872-6.
  10. Brown T, Ghelani AA, Yeung D, Nguyen HB. Comparative effectiveness of physician diagnosis and guideline definitions in identifying sepsis patients in the emergency department. J Crit Care 2015;30(1):71-7.
  11. Mackenzie CF, Gao C, PF Hu, Anazodo A, Chen H, Dinardo T, et al. Comparison of de- cision-assist and clinical judgment of experts for prediction of lifesaving interven- tions. Shock 2015;43(3):238-43.
  12. Brunauer A, Kokofer A, Bataar O, Gradwohl-Matis I, Dankl D, Bakker J, et al. Changes in Peripheral perfusion relate to visceral organ perfusion in early septic shock: a pilot study. J Crit Care 2016;35:105-9.
  13. Heponiemi T, Aalto AM, Pekkarinen L, Siuvatti E, Elovainio M. Are there high-risk groups among physicians that are more vulnerable to On-call work? Am J Emerg Med 2015;33(5):614-9.
  14. Visser A, Wolthuis A, Breedveld R, ter Avest E. HEART score and clinical gestalt have similar diagnostic accuracy for diagnosing ACS in an Unselected population of pa- tients with chest pain presenting in the ED. Emerg Med J 2015;32(8):595-600.
  15. Nugus P, Holdgate A, Fry M, Forero R, McCarthy S, Braithwaite J. Work pressure and patient flow management in the emergency department: findings from an ethno- graphic study. Acad Emerg Med 2011;18(10):1045-52.
  16. Knapp J, Bernhard M, Hainer C, Sikinger M, Brenner T, Schlechtriemen T, et al. Is there an association between the rating of illness and injury severity and the experience of emergency medical physicians? Anaesthesist 2008;57(11): 1069-74.
  17. Venkatesh AK, Slesinger T, Whittle J, Osborn T, Aaronson E, Rothenberg C, et al. Preliminary performance on the new CMS Sepsis-1 National Quality Measure: early insights from the emergency quality network (E-QUAL). Ann Emerg Med 2017. https://doi.org/10.1016/j.annemergmed.2017.06.032 [Epub ahead of print].
  18. Guyette FX, Meier EN, Newgard C, McKnight B, Daya M, Bulger EM, et al. ROC inves- tigators: a comparison of prehospital lactate and systolic blood pressure for predicting the need for resuscitative care in trauma transported by ground. J Trauma Acute Care Surg 2015;78(3):600-6.
  19. Jafar AJ, Junghans C, Kwok CS, Hymers C, Monk KJ, Gold E, et al. Do physiological scoring and a novel point of care metabolic screen predict 48-h outcome in admis- sions from the emergency department resuscitation area? Eur J Emerg Med 2016; 23(2):130-6.
  20. Cevik AA, Dolgun H, Oner S, Tokar B, Acar N, Ozakin E, et al. elevated lactate level and shock index in nontraumatic hypotensive patients presenting to the emergency de- partment. Eur J Emerg Med 2015;22(1):23-8.
  21. Gillespie I, Rosenstein PG, Hughes D, et al. Update: clinical use of plasma lactate. Vet Clin North Am Small Anim Pract 2017;47(2):325-42.
  22. Andersen LW, Mackenhauer J, Roberts JC, Berg KM, Cocchi MN, Donnino MW, et al. Etiology and therapeutic approach to elevated lactate levels. Mayo Clin Proc 2013; 88(10):1127-40.
  23. Gregory SD, Cooney H, Diab S, Anstey C, Thom O, Fraser JF. In vitro evaluation of an ultrasonic cardiac output monitoring (USCOM) device. J Clin Monit Comput 2016; 30(1):69-75.
  24. O’Loughlin E, Ward M, Crossley A, Hughes R, Bremner AP, Corcoran T. Evaluation of the utility of the Vigileo FloTrac((TM)), LiDCO((TM)), USCOM and CardioQ((TM)) to detect hypovolaemia in conscious volunteers: a proof of concept study. Anaesthesia 2015;70(2):142-9.

Leave a Reply

Your email address will not be published. Required fields are marked *