Variations and Tendencies in Clinical Decision-Making of MDH Hospital Staff

in Patients with Acute Nontraumatic RLQ Abdominal Pain

Reynaldo O. Joson, MD, MHA, MHPEd, MS Surg

Overall Coordinator

Members of Task Force on Clinical Decision-Making on RLQ Abdominal Pain*

Drs. R. Aman, L. Bernardo, J. Caraos, T. Catabas, G. Cruz, M. Dalluay, M. de Leon,

A. Dumlao, R. Dy, V. Mag-iba, K. Reubenfeldt, and E. San Luis

Introduction

In 1998, the Manila Doctors Hospital created a quality assurance committee to improve management of patients with acute nontraumatic right lower quadrant (RLQ) abdominal pain.

One of the committee’s strategies is to develop an evidence-based clinical practice guideline.

The conceptual framework of the guideline development project is as follows:

Input

Throughput

Output

Variations in diagnostic and treatment process Confused patients Confused medical students Costly diagnostic and treatment process Delays in diagnosis and treatment Missed diagnosis Wrong diagnosis Medicolegal suits

evidence-based clinical practice guidelines on acute nontraumatic right lower quadrant abdominal pain

Standardized diagnostic process and treatment process Well-informed patients Well-informed students Cost-effective diagnostic and treatment process Efficient diagnosis and treatment Reduced missed diagnosis Reduced wrong diagnosis Reduced medicolegal suits -------------------------------------- Decrease negative appendectomy rate to an acceptable level without significantly increasing perforation rate and its morbidity and mortality consequences

Various task forces were created for the documentation of baseline data as well as for the formulation of clinical practice guidelines. These task forces are working simultaneously and in a concerted cooperative manner.

This paper is a report of the task force assigned to gather data on the present practice or manner of decision-making of MDH medical staff on clinical diagnosis and use of ancillary diagnostic procedures in patients with acute nontraumatic RLQ abdominal pain.

General Objective:

To gather data on the present practice or manner of decision-making of MDH

medical staff on clinical diagnosis and use of ancillary diagnostic procedures in

patients with acute nontraumatic RLQ abdominal pain.

Specific Objectives:

In the decision-making on clinical diagnosis and use of ancillary diagnostic procedures

in patients with acute nontraumatic RLQ abdominal pain,

To document:

1. Overall variations, if present.

2. Variations, if present, among the different levels of medical staff.

3. Variations, if present, among the different specialties.

4. Overall tendencies.

Delimitation of Paper

This paper was delimited to documenting variations and tendencies in the decision-making on clinical diagnosis and use of ancillary diagnostic procedures in patients with acute nontraumatic RLQ abdominal pain.

Significance of the Study

The findings in this paper would be correlated with published research evidences that will be gathered in a separate project. The findings will certainly play a role in the formulation of evidence-based clinical practice guidelines to be used in Manila Doctors Hospital in providing quality health care to patients with RLQ abdominal pain.

Methodology

Various clinical scenarios were created which included two genders, two age groups (5-4 and 15-50 years old), and various combinations of signs and symptoms. (Note: a third age group >50 years old was included in the original plan; however, this was not administered due to lack of time and cooperative respondents.) Thus, there were a total of 4 sets of scenarios: male aged 5-14 set; female aged 5-14 set; male aged 15-50 set; and female aged 5-50 set. There were 23 scenarios per set. See Appendix 1 for samples of clinical scenario.

The clinical scenarios tried to elicit a respondent’s decision-making in clinical diagnosis, degree of certainty of diagnosis, and use of ancillary diagnostic procedures.

They were constructed in such a manner as to bring out any weight that the respondents might possibly place on shifting pain, rebound tenderness, guarding, and associated dysuria, diarrhea, vaginal discharge, and pregnancy.

A questionnaire was also formulated to be answered together with the clinical scenarios. It tried to elicit the approach of respondents in the clinical evaluation of patients with RLQ pain, such as the weights they might place on certain set of signs and symptoms with greatest likelihood of appendicitis, their usual physical examination maneuvers, and their practice of requesting ancillary diagnostic procedures.

A sample of MDH medical staff with different levels of training and different specialties were selected to be the respondents. Initially, there was an attempt for a randomization of responders. However, when difficulty was encountered in the forms of unavailability and uncooperativeness, choice of respondents was eventually based on convenience.

Essentially, what was done was a survey, therefore, the research design was a descriptive one.

To determine the presence of variations and the tendencies in the clinical decision-making, measures of central tendencies and dispersions were used.

Results

I. Concurrence and Variations

Table 1 shows the number of respondents, distribution as to specialties and consultants, fellows, residents, and interns and scenarios responded to. The singular gastroenterology fellow was classified together with the gastroenterology consultants.

Table 1. Respondents and scenarios responded to.

Respondents	Scenarios
	Males		Females
	5-14	15-50	5-14	15-50
Surgery
Consultants	3	3	3	3
Residents	3	3	3	3
Internal Medicine
Consultants/fellow - Gastro		3		3
Residents - General		5		5
Pediatrics
Consultants	5		5
Residents	5		5
Obstetrics-Gynecology
Consultants				5
Residents				5
Emergency Room Consultants		4		4
Interns		10		10
Total = 98	16	28	16	38

With some respondents answering more than one set of scenarios, there was a total of 98 sets answered by 41 consultants, 37 residents, and 20 interns vs the actual number of person-respondents of 20, 21, and 20 respectively. (Table 2)

Table 2. Respondents by type: consultants, residents, and interns.

Respondents by Types (actual no. of person)	Scenarios
	Males		Females
	5-14	15-50	5-14	15-50	Total
Consultants (20)	8	10	8	15	41
Residents (21)	8	8	8	13	37
Interns (20)	0	10	0	10	20
Total	16	28	16	38	98

Five different specialties were included in the sampling. The male and female age 5-14 sets were answered by surgeons and pediatricians. The female aged 15 50 set was answered by surgeons, internists, obstetrician-gynecologists, and emergency department physicians. For the male aged 15-50 set, surgeons, internists, and emergency department physicians. (Table 3)

Table 3. Respondents by specialties.

Respondents by Specialty	Scenarios
	Males		Females
	5-14	15-50	5-14	15-50
Surgery	6	6	6	6
Internal Medicine		8		8
Pediatrics	10		10
OB-GYN				10
Emergency Department		4		4

As shown in Table 4, there was an overall marked variation in the giving of a primary clinical diagnosis of acute appendicitis given the same set of scenarios. The overall range of concurrence was 7 to 100 with a mean of 60. The mean and median percentages of concurrence were 59 and 63 respectively.

There were multimodal values with a low 38% to a high 88% of concurrence.

Table 4. Degree of overall concurrence (variation) in the primary clinical diagnosis of acute appendicitis.

Scenario Set		Range [d] (%)	Mean (%)	Median (%)	Mode (%)
Male	5-14	38-100 [62]	71	81	38,81,88
Female	5-14	13-100 [87]	61	63	63,69,75
Male	15-50	18- 93 [75]	58	57	57,75
Female	15-50	7- 93 [86]	47	50	68
Overall		7-100 [60]	59	63	MULTIMODAL

Overall, the mean range of certainty of a clinical diagnosis was 33% with an actual range of 0-38%. The mean and median percentages of certainty were only 12 % and 10% respectively. In other words, in about 90% of the time, the respondents were not certain of their clinical diagnosis to justify the need for ancillary diagnostic procedures. (Table 5)

Table 5. Degree of overall concurrence (variation) in the certainty of the primary clinical diagnosis.

Scenario Set		Range [d] (%)	Mean (%)	Median (%)
Male	5-14	0-38 [38]	18	19
Female	5-14	0-37 [37]	8	6
Male	15-50	0-25[25]	11	7
Female	15-50	0-32[32]	11	7
Overall (Average)		0-38[33]	12	10

Overall, there was a marked variation in the use of ancillary diagnostic procedures given the same set of scenarios, with a mean of 66% concurring in the use of combination of CBC and urinalysis. (Table 6)

Table 6. Degree of overall concurrence (variation) in the use of ancillary diagnostic procedures.

Ancillary Diagnostic Procedures	Mean (%)
CBC and Urinalysis	66
Observation	4
Others CBC alone, urinalysis alone, fecalysis, plain abdomen, barium enema, ultrasound, etc.	30

Comparing the degree of concurrence or variation among the different levels of medical staff in the primary diagnosis of acute appendicitis, among the consultants, there was a 66% concurrence; among residents, 55%; interns, 45% . Although the percentage of concurrence was highest among the consultants as compared to that of the residents and interns, the 66% by itself constituted a marked variation in the consultants’ giving of a primary clinical diagnosis of acute appendicitis given the same set of scenarios. (Table 7).

Table 7. Degree of concurrence (variation) among and between different levels of medical staff in the primary diagnosis of acute appendicitis.

Scenario Set		Consultants (median %)	Residents (median %)	Interns (median %)
Male	5-14	88	62	na
Female	5-14	75	62	na
Male	15-50	50	50	60
Female	15-50	53	46	30
Average		66	55	45

As shown in Table 8, the surgeons had the highest degree of concurrence (80 and 81%). Comparing the pediatricians and the surgeons, their percentage of concurrence in the diagnosis of acute appendicitis given the same set of scenarios were not markedly different, 76% vs 81%. However, comparing the surgeons, obstetric-gynecologists, emergency department physicians, and internists, given the same set of scenarios, the percentage of concurrence of the nonsurgeons was markedly different from that of the surgeons.

Table 8. Degree of concurrence (variation) among and between the different specialties in the primary clinical diagnosis of acute appendicitis.

Scenario Set		Pediatricians (median %)	Surgeons (median %)
Male	5-14	88	83
Female	5-14	64	78
Average		76	81

Scenario Set		Surgeons (median %)	Ob-Gynecologists (median %)	ED Physicians median %)	Internists (median %)
Male	15-50	76	na	62	46
Female	15-50	83	60	50	na
Average		80	60	56	46

II. Tendencies

In the absence of urinary and bowel disturbance, vaginal discharge and suspicion of pregnancy, in all age groups with acute nontraumatic RLQ pain and tenderness, majority of the respondents considered presence of direct RLQ tenderness with rebound and guarding as having the greatest likelihood of having acute appendicitis. Next was direct tenderness with rebound followed by direct tenderness with guarding. Mere rebound without direct tenderness and guarding was considered to have the least likelihood. (Table 9)

Table 9. RLQ physical signs: positive likelihood for acute appendicitis. (Questionnaire)

Greatest -----Likelihood of Acute Appendicitis ---- Least

Set of RLQ physical signs	1	2	3	4	5
Direct tenderness only				x
Direct tenderness, rebound		x
Direct tenderness, rebound, guarding	x
Direct tenderness, guarding			x
Rebound tenderness only					x

However, looking at the clinical scenarios (Tables 10 and 11), there were discrepancies on the differential weights placed on guarding and rebound (Table 10) and on direct and rebound tenderness (Table 11).

In the clinical scenarios, there was a tendency to put more weight on guarding than on rebound (65% vs 54%). In the questionnaire, more weights were placed on rebound tenderness.

In the clinical scenarios, there was a tendency to put more weight on rebound tenderness than on direct tenderness (44% vs 41%). In the questionnaire, more weights were placed on direct tenderness.

What was consistent in both clinical scenarios and questionnaire were the following:

1. More weights were placed on the presence of a combination of tenderness as compared to a single kind of tenderness.

2. More weights were placed on three kinds of tenderness, namely, direct, rebound, and guarding as compared to two kinds of tenderness.

Table 10. Weights placed on the different kinds of RLQ tenderness for a diagnosis of acute appendicitis. (Clinical Scenarios)

Scenario Set		Direct Tenderness (median %)	Direct Tenderness Rebound (median %)	Direct Tenderness Guarding (median %)	Direct Tenderness Rebound Guarding (median %)
Male	5-14	46	65	82	86
Female	5-14	44	50	69	79
Male	15-50	23	65	64	67
Female	15-50	26	36	44	42
Average		35	54	65	68

Table 11. Weights placed on direct and rebound tenderness in the diagnosis of acute appendicitis in the absence of associated symptoms and signs. (Clinical Scenarios)

Scenario Set		Direct Tenderness (median %)	Rebound Tenderness (median %)	Direct and Rebound Tenderness (median %)
Male	5-14	38	38	88
Female	5-14	50	44	63
Male	15-50	29	50	75
Female	15-50	46	43	75
Average		41	44	75

In the clinical scenarios, there was a tendency to put positive weights on epigastric or periumbilical pain shifting to the RLQ for a diagnosis of acute appendicitis. On the other hand, there was a tendency to put negative weights on LLQ pain shifting to the RLQ. (Table 12).

Table 12. Weights placed on shifting prodromal pain for a diagnosis of acute appendicitis. (Clinical Scenarios)

Scenario Set		RLQ (median %)	Epigastric or periumbilical to RLQ (median %)	LLQ to RLQ (median %)
Male	5-14	41	74	41
Female	5-14	75	82	70
Male	15-50	89	86	79
Female	15-50	81	82	56
Average		71	81	62

In the clinical scenarios, in the presence of symptoms like diarrhea, dysuria, vaginal discharge, and amenorrhea which could serve as cues for conditions other than acute appendicitis, overall, there was a tendency to consider strongly a nonappendicitis diagnosis. (see horizontal comparison of cells in the Table 13) However, there was a decreasing tendency to consider a nonappendicitis diagnosis and corollary to this, a increasing tendency to consider appendicitis, in the presence of rebound tenderness and guarding. (see vertical comparison of cells in the Table 13).

Table 13. Weights placed on presence of associated symptoms in the clinical diagnosis of acute appendicitis in patients with acute RLQ pain. (Clinical Scenarios)

Male Patients 5-14 and 15-50
Kinds of RLQ Tenderness	No associated symptoms (median %)	w/ Diarrhea (median %)	w/ Dysuria (median %)
Direct Tenderness	33	37	33
Direct Tenderness / rebound or guarding	82	71	56
Direct Tenderness/rebound/guarding	88	78	69

Female Patients 5-14 and 15-50
Kinds of RLQ Tenderness	No associated symptoms (median %)	w/ Vaginal Discharge (median %)	Pregnant Suspect (median %)
Direct Tenderness	48	12	45
Direct Tenderness / rebound or guarding	75	25	50
Direct Tenderness/rebound/guarding	91	41	50

In the questionnaire, overall, there was a tendency to routinely elicit cough, Rovsing, psoas and obturator signs and perform rectal examination after eliciting presence of RLQ tenderness. However, there was a tendency not to routinely perform internal examination in females. (Table 14) Comparing the different specialists, there is a tendency for the surgeons not to routinely do all kinds of maneuvers in the physical examinations.

Table 14. Tendency for routine performance of certain physical examination maneuvers after eliciting presence of RLQ tenderness.

Maneuver	Pedia (%)	Internists (%)	Ob-Gyne (%)	Surgeons (%)	Overall (%)
Cough sign	80	100	100	50	83
Rovsing sign	80	100	100	70	88
Psoas sign	80	100	100	50	83
Obturator sign	80	100	100	50	83
Rectal exam	80	100	100	50	83
Internal exam	60	63	55	0	45
Mode	80	100	100	50	83

In the questionnaire, overall, there was a tendency to routinely request for CBC and urinalysis, but not ultrasound. (Table 15) Comparing the different specialties, there is a tendency for the surgeons not to routinely request for CBC and urinalysis.

As to routinely observing and monitoring the abdominal findings, overall, 95% of respondents claimed they would do it. However, in the clinical scenarios, observation as an ancillary diagnostic procedure was chosen in only 4% of cases. (Table 6)

Table 15. Tendency for routinely requesting for ancillary diagnostic procedures after eliciting presence of RLQ tenderness.

Examination	Pedia (%)	Internists (%)	Ob-Gyne (%)	Surgeons (%)	Overall (%)
CBC	100	100	100	50	88
Urinalysis	100	100	77	50	82
Mean	100	100	89	50
Ultrasound	50	25	33	40	37
Observation	80	100	100	100	95

Summary of Findings

1. Variations

There were marked variations overall; between and among consultants, residents, and interns; and between and among different specialties in the clinical diagnosis and use for ancillary diagnostic procedures.

2. Tendencies

There were tendencies toward the following:

1. In the absence of cues for an alternative nonappendicitis diagnosis, placing greatest weight on a combination of direct tenderness, rebound tenderness, and guarding for a clinical diagnosis of acute appendicitis. Corollary to this, placing more weights on the presence of two or more kinds of tenderness than on a single kind.

2. Using shifting pain from epigastric or periumbilical area to the RLQ as an important cue in the clinical diagnosis of acute appendicitis. Corollary to this, nonreliance on shifting LLQ to RLQ as a cue for the clinical diagnosis of acute appendicitis.

3. In the presence of cues for an alternative nonappendicitis condition, not to consider strongly an appendicitis diagnosis. However, in the presence of rebound tenderness and guarding, to consider strongly an appendicitis diagnosis together with the nonappendicitis one.

4. To routinely elicit cough, Rovsing, psoas, and obturator signs and perform rectal examination. Not to routinely perform internal examination in patients with RLQ pain.

5. Uncertainty in clinical diagnosis in the majority of cases that would prompt request for ancillary diagnostic procedure.

6. To routinely perform ancillary diagnostic procedure, especially CBC and urinalysis.

7. For surgeons not to routinely do all kinds of maneuvers in the physical examination and request for CBC and urinalysis.

8. Not to perform observation as an ancillary diagnostic procedure.

9. No clearcut trend on where to put more weights on: direct or rebound tenderness; rebound tenderness or guarding.

Discussion

Up to this day and age, more than 100 years since Dr. Reginald Fitz first formally described acute appendicitis, there are still a lot of scientific papers on how best to evaluate patients with acute nontraumatic RLQ abdominal pain. The spectrum of the topics ranges from the basic tools of history and physical examination to the most sophisticated and advanced technology such as ultrasound, computed tomography, and laparoscopy. The sad thing is that all the studies have not come out with reliable and precise recommendations and guidelines. The studies often conflict with each other. Some would still advocate traditional investigative tools while others, advanced technologies. Some would rely on anorexia, some would not, for the diagnosis of appendicitis. Some would rely on rebound tenderness while some would rely on guarding as a sign of peritonitis. All these just show that the management of patients with acute nontraumatic RLQ abdominal pain continues to be very challenging and confusing.

Our study shows that there is a marked variation in the clinical and ancillary diagnostic processes utilized by the staff of Manila Doctors Hospital. This is not surprising as this is the same situation globally.

The purpose of this paper is just to document the presence of marked variation to justify the need for the development of evidence-based clinical practice guidelines.

Tendencies on the clinical and paraclinical diagnostic processes were identified. Identification of the tendencies will be of great help to the developers of the clinical practice guidelines. The developers can initially focus on looking for evidences to support the tendencies. If in the process of looking for evidences to support the tendencies, evidences turn out not to favor the tendencies, then recommendations should be given to this effect.

This paper is in no position to comment on the tendencies of the medical staff. Whether there should be a redirection or reengineering of the present tendencies will be answered after the development of evidence-based clinical practice guideline is completed.

An example of an evidence which may be helpful in resolving the issue on which symptoms and signs should be given reliance on for a diagnosis of acute appendicitis is the meta-analysis of Wagner et al. (14) It is a study of more than 4,000 patients with the likelihood ratios for important clinical diagnostic cues tabulated.

Procedure	Sensitivity	Specificity	LR+ (95%CI)	LR-(95% CI)
RLQ pain	0.81	0.53	7.31-8.46	0-0.28
Rigidity	0.27	0.83	3.76(2.90-4.78)	0.82(O.79-0.85)
Migration	0.64	0.82	3.18(2.41-4.21)	0.52(0.42-0.59)
Pain before vomiting	1.00	0.64	2.76(1.94-3.94)	NA
Psoas sign	0.16	0.95	2.38(1.21-4.67)	0.90(0.83-o.98)
Fever	0.67	0.79	1.94(1.63-2.32)	0.58(0.51-0.67)
Rebound tenderness test	0.63	0.69	1.10-6.30	0-0.86
Guarding	0.74	0.57	1.65-1.78	0-0.54
No similar pain previously	0.81	0.41	1.50(1.36-1.66)	0.32(0.25-0.42)
Rectal tenderness	0.41	0.77	0.83-5.34	0.36-1.15)
Anorexia	0.68	0.36	1.27(1.16-1.38)	0.64(0.54-0.75)
Nausea	0.58	0.37	0.69-1.20	0.70-0.84)
Vomiting	0.51	0.45	0.92(0.82-1.04)	1.12(0.95-1.33)

In this table, one can see that rigidity and migration with a likelihood ratios of more than 3 could be recommended as cues for a clinical diagnosis of acute appendicitis.

Comparing rebound tenderness and guarding, it is difficult to make any conclusions as to which one is more reliable because of the wide range for the former sign.

Conclusion

The data-gathering showed marked variations in the clinical decision-making of medical staff of Manila Doctors Hospital which would justify the development of an evidence-based clinical practice guideline. Hopefully, this guideline will not only resolve the variations but will reduce the confusion among patients and health care providers alike. The ultimate impact of the guideline will be a better health care outcome.

References

1. Anatol TI, Holder Y. A scoring system for use in the diagnosis of acute abdominal pain in childhood. West Indian Med J 1995; 44:2, 67-9.

2. Balsano N, Cayten CG. Surgical emergencies of the abdomen. Emerg Med Clin North Am 1990;8:399-410.

3. Bugliosi TF, Meloy TD, Vukoy LF. Acute abdominal pain in the elderly. Ann Emerg Med 1990;19:1383-1386.

4. Dixon JM, Elton RA, Rainey JB, Macleod DAD. Rectal examination in patients with pain in the right lower quadrant of the abdomen. BMJ 1991;302:386-388.

5. Fenyo G. Acute abdominal disease in the elderly. Am J Surg 1982;143:751-754.

6. Hawthorn IE. Abdominal pain as a cause of acute admission to hospital. J R Coll Surg Edinb 1992; 37:389-93.

7. Irvin TT. Abdominal pain: a surgical audit of 1190 emergency admissions. Br J Surg 1989;76:1121- 1124.

8. Putnam TC, Gagliano N, Emmens RW. Appendicitis in children. Surg Gynecol Obstet 1990;170:527- 532.

9. Simmen HP, Decurtins M, Rotzer A, Duff C, Brutsch HP, Largiader F. Emergency room patients with abdominal pain unrelated to trauma: prospective analysis in a surgical university hospital. Hepatogastroenterology. 1991 38:4, 279-82.

10. Wilson DH, Wilson PD, Walmsley RG, Horrocks JC, De Dombal FT. Diagnosis of acute abdominal pain in the accident and emergency department. Br J Surg 1977; 64:250-4.

11. Institute of Medicine (1990). Clinical practice guidelines: directions for a new program. (eds. Field, MJ and Lohr, KN). Institute of Medicine, National Academy Press, Washington, DC.

12. Quality of Care and Health Outcomes Committee. Office of the National Health and Medical Research Council (Australia). Guidelines for the development and implementation of clinical practice guidelines. October, 1995.

13. Fitz RH. Perforating inflammation of the vermiform appendix with special reference to its early diagnosis and treatment. Am J Med Sce 1886;92:321-346.

14. Wagner JM, McKinney WP, Carpenter JL. Does this patient have appendicitis? JAMA 1996;275:1589-1594.

Acknowledgment:

*Members of Task Force on Clinical Decision-Making on RLQ Abdominal Pain

MDH Interns Batch 1998-1999

Rowena Aman, MD

Luzviminda Bernardo, MD

Joselino Caraos, MD

Tomas Catabas, MD

Gary Cruz, MD

Melvin Dalluay, MD

Melissa de Leon, MD

Ariel Dumlao, MD

Renelyn Dy, MD

Vivien Mag-iba, MD

Karen Reubenfeldt, MD

Emil San Luis, MD

Appendix 1

Samples of the Clinical Scenarios

1. Male patient, 15-50 years old

Clinical Findings and Data (with 2 physical exams within 30 minutes)

RLQ PAIN AND TENDERNESS

Onset < 7 days

Started in and confined to RLQ

Direct RLQ tenderness, NO rebound, NO guarding

No bowel or urinary disturbance

No other associated symptoms and signs

2. Female patient, 5-15 years old

Clinical Findings and Data (with 2 physical exams within 30 minutes)

RLQ PAIN AND TENDERNESS

Onset < 7 days

Started in and confined to RLQ

Direct RLQ tenderness, with rebound and guarding

With diarrhea

No bowel or urinary disturbance

No other associated symptoms and signs

Clinical Decision-making on:

Primary Clinical Diagnosis:

Secondary Clinical Diagnosis:

Degree of Certainty of Clinical Diagnosis:

Certain that there is no need for ancillary diagnostic procedure

Uncertain that there is need for ancillary diagnostic procedure

Ancillary Diagnostic Procedure (observation included):