05 Screening and diagnostic tests 3.pptx

Screening and diagnostic

1. School of Public Health Epidemiology and Health Statistics Name: Li Yang Phone: [phone] QQ: 527293391 Wechat: liyang527293391 self-introduction

2. Preventive medicine Chapter VI -- Screening and diagnostic tests Teaching object: Class 2022 Clinical International class Class Time: November 10, 2024 (Sunday 6-8 sessions) Location: Room 305, Building 7

3. Teaching requirement 【 Knowledge objective 】 : 1. Can explain the concept of diagnostic tests; The concept of normative method (gold standard); 2. Can understand and distinguish the evaluation indicators of diagnostic tests: sensitivity, specificity, false positive rate (misdiagnosis rate), false negative rate (missed diagnosis rate) concept and significance; 3. Can distinguish the relationship between sensitivity and specificity; 4. Can understand the concept and significance of positive predictive value and negative predictive value; 5. Can explain and understand the concept and use of parallel test and series test. 6. Can list the diagnostic criteria of cases; 7. Can name the type of diagnostic indicators;

4. 【 Capability objective 】 : Can be analyzed using evaluation methods of screening and diagnostic tests. Be able to understand the diagnostic criteria (gold criteria) of cases. 【 Emotional goals 】 : To be able to judge and experiment the value of scientific methods to cultivate students' scientific spirit. Teaching requirement

5. S ection 1 ： S ummarize Section 2: Evaluation of diagnostic tests and screening tests Section 3 : the method of improving the test efficiency content

6. 一、Concept screening is the use of quick and easy tests, examinations, or other measures to detect in a healthy population those who appear healthy but are suspected of having a disease or defect . screening is also called screening, and the various means and methods used in screening are called screening tests, which can be physical examinations, laboratory examinations and questionnaires.

7. Purpose of screening 1. Secondary prevention of diseases, early detection, early diagnosis and early treatment. Screening for breast cancer. 2. Primary prevention of diseases, high-risk groups, take corresponding intervention measures. For example, screening for high blood pressure to prevent stroke . 3. Identify the early stages of the disease to help understand the natural history of the disease and reveal the "iceberg phenomenon" of the disease. 4. Rational allocation of health resources .

8. It refers to the basic understanding and judgment of the patient's condition after the medical staff collects various information through detailed examination and investigation. Accordingly, the next treatment or intervention plan should be formulated to complete the treatment task. The various tests and investigation methods used for diagnosis are called diagnostic tests. The meaning of diagnostic test is extensive, it includes a variety of laboratory examinations, including radiation, B-ultrasonography and other physical examination means, but also includes medical history, physical examination and other clinical data. diagnosis

9. To make timely and correct judgment of the patient's condition, so as to take corresponding and effective treatment measures. Diagnostic tests can also be used to follow up cases, determine the outcome of the disease, judge the efficacy and estimate the prognosis, and monitor the side effects of treatment. Purpose of diagnosis

10. 1. The disease or defect being screened is a major local health problem. 2. Methods and conditions for further diagnosis of the disease or defect being screened. 3. There are conditions for effective treatment and intervention for found and diagnosed patients and high-risk groups, and there should be uniform standards. 4. Know the natural history of the disease being screened and have early signs and symptoms that can be identified or measured. 5. The screening test should be quick, simple, economical, reliable, safe, effective and easy to be accepted by the public. Principles of application of screening

11. Diagnosis is the premise of clinical effective treatment measures, so the diagnostic test must be able to correctly judge patients and non-patients, but also accepted by the client, so the diagnostic test requirements : 1. High sensitivity and specificity. 2. Fast, simple, cheap and easy to carry out. 3. Safe, reliable, minimize patient injury and pain.

12. Difference between diagnostic test and screening test Screening test Diagnostic test aim To distinguish suspicious patients from those who may not be sick It's used to distinguish sick people from people who are suspected to be sick but aren't object A person who is healthy or apparently healthy Sick or suspicious patients demand Fast, simple and highly sensitive to detect all possible patients High accuracy, specificity and sensitivity to diagnose all patients and exclude all non-patients as much as possible expense Simple, cheap way Generally higher price handle Further diagnostic tests are required to confirm the positive results Those who are positive need further observation and prompt treatment

13. Whether it is a "diagnostic test" or a "screening test", it is necessary to evaluate the authenticity and value of the "diagnosis" of the disease.

14. 一、The basic steps of test evaluation The evaluation of the trial is to compare the results of the trial to be evaluated with the results of the "gold standard" - synchronous blind method to determine the authenticity and value of the method for the "diagnosis" of the disease. Section II Evaluation of screening test

15. Gold standard Specific population Patient nonpatient Method to be evaluated Evaluation index Figure 1 Basic steps of test evaluation ＋ － ＋ －

16. （一）Setting the 'gold Standard' "Gold standard" refers to the current clinical medical community recognized as the most reliable, the most authoritative, can reflect the actual situation of disease or disease is called the gold standard, is the most reliable way to diagnose disease, also known as standard diagnosis. Alive/Autopsy Surgical discovery Microbial culture Special inspection Image diagnosis Long-term follow-up results

17. （二）Selection of research objects Study subjects used to evaluate diagnostic and screening tests should be representative of the target population to which screening and diagnostic tests are likely to be applied. Therefore, the study subjects for screening should be selected from the community population , and the study subjects for diagnostic tests should be selected from the suspicious or high-risk population. The "sick" cases confirmed by the gold standard should include typical and atypical cases, early, middle and late cases, mild, moderate and severe cases, and patients with and without complications. Representativeness gives the evaluation results universal significance. The results of the trial evaluation from the male and young population are generally not applicable to the female and elderly population.

18. （三）Calculation of sample size Sensitivity of the test to be evaluated Specificity of the test to be evaluated Significance test level α , generally 0.05 The allowable error δ is generally 0.05 ~ 0.10 When the sensitivity and specificity are close to 50%, the approximate formula can be used Sensitivity or specificity of the method to be evaluated This formula applies when p is close to 50%, and when the expected sensitivity or specificity is less than 20% or greater than 80%, the estimated sensitivity or specificity needs to be converted by square root arcsine.

19. （四）Synchronous blind test The case group and the non-case group identified by the gold standard were tested by synchronous blind method with the trial to be evaluated.

20. （五）Sort out and analyze data Evaluation test Gold standard diagnosis total case noncase positive a ( True positive ) b ( False positive ) a+b n egative c ( F alse negative ) d ( True negative ) c+d total a+c b+d n Table 1 Adjustment table of evaluation test

21. In the collection and analysis of test data, the use of blind method is of great significance, which can reduce the subjective bias . Assuming that blindness is not applied, researchers may be inclined to judge gold-confirmed patients as positive and gold-diagnosed non-patients as negative for equally dubious results, thus overestimating the value of the trial to be evaluated. （六）Quality control

22. In addition to the use of synchronous blind testing to ensure the authenticity of the results, the test instrument model, test conditions, test methods, the quality of reagents used, labels, etc. should be unified and standardized, objective indicators should be used as far as possible, and investigators should be strictly trained to minimize the error .

23. 二、Test evaluation index （一） A uthenticity (Validity) Refers to the degree to which the measured value conforms to the actual value (gold standard), so it is also called accuracy . S ensitivity and F alse negative rate S pecificity and F alse positive rate Youden's index C rude agreement

24. sensitivity Also known as sensitivity or true positive rate , refers to the percentage of cases confirmed by the gold standard who are also judged positive by the evaluation test. It may reflect the ability of the test to be evaluated to correctly identify cases of actual disease. The ideal value should be 100%.

25. specificity Also known as the true negative rate , refers to the percentage of non-cases confirmed by the gold standard who are also judged negative by the evaluation test. It can reflect the ability of the test to be evaluated to correctly determine that the subjects who actually do not have a disease. The ideal value should also be 100%.

26. False negative rate Also known as the missed diagnosis rate , refers to the percentage of cases confirmed by the gold standard to be evaluated and wrongly judged negative by the test, the ideal value should also be 0%, the M issed diagnosis rate =1- S ensitivity. The higher the sensitivity, the lower the missed diagnosis rate.

27. False positive rate A lso known as misdiagnosis rate, refers to the percentage of non-cases diagnosed by gold standard that are misdiagnosed as positive by tests to be evaluated. The ideal value should be 0%, and the misdiagnosis rate =1- specificity . The higher the specificity, the lower the misdiagnosis rate.

28. Because many tests use quantitative or semi-quantitative methods to determine positive and negative, and patient and non-patient measurement values can not be completely separated, there is overlap , so sensitivity and specificity often lose one or the other. Often, the sensitivity of the test is improved but the specificity of the test is reduced , while the specificity of the test is improved but the sensitivity of the test is reduced. The following figure shows the relationship between the two . When the judgment criterion is moved to B , the specificity is improved and the sensitivity is reduced; when the judgment criterion is moved to A , the sensitivity is improved and the specificity is reduced.

29. Figure 2 Distribution of serum calcium levels in the population People with hyperthyroid gland Normal population Serum calcium level frequency

30. Jorden index Also known as the C orrect index Is the sum of sensitivity and specificity minus 1 Represents the total ability of the test method to detect real patients and non-patients The range is between 0 and 1, with the higher the index, the higher the truth Jorden index = （ S ensitivity +S pecificity） -1 =1- （ False negative rate + False positive rate）

31. Rough consistency Refers to the sum of the number of true positive and true negative cases detected in the test as a percentage of the number of subjects. It reflects the degree of agreement between the test results and the gold standard diagnostic results.

32. Table 2 Relationship between disease status and test results (virtual data) It indicates that 60.00% of patients with confirmed hyperparathyroidism have positive or abnormal serum calcium tests.

33. It indicated that 95.00% of the patients without hyperparathyroidism had negative or normal serum calcium tests. This means that 40.00% of patients with confirmed hyperparathyroidism have a negative or normal serum calcium test. It indicated that 5.00% of the patients without hyperparathyroidism had positive or abnormal serum calcium tests.

34. Jorden index = （ S ensitivity +S pecificity） -1 =1- （ False negative rate + False positive rate） Jorden index = （ 0.60+0.95 ） -1=0.55 =1- （ 0.4+0.05 ） =0.55 The probability that both positive and negative results of the serum calcium test are correct is 86.25%.

35. （二） Evaluation of test reliability Also known as reliability, precision, or repeatability , refers to the degree of stability that the same result is obtained when the same subject is repeatedly measured with a measurement tool (such as the test to be evaluated) under the same conditions. Factors affecting test reliability Biological differences among the subjects themselves Observer difference Differences in test methods

36. Index to evaluate test reliability Standard deviation Coefficient of variation (CV) agreement/consistency rate Kappa value

37. The coincidence rate is also called the consistency rate It is the proportion of the total number of patients that the results of the screening test are the same as those of the standard diagnosis Table 3 Data summary of reliability evaluation

38. However, it is not comprehensive to evaluate the consistency of screening or diagnostic tests by coincidence rate alone. First, the coincidence rate does not indicate the specific cause of the inconsistency: are the positive and negative results the same for both observers? Or was one of the observers more likely to be positive? Second, a lot of consistency is caused by chance, but the coincidence rate does not indicate the magnitude of chance. To consider the influence of chance factors on observation consistency, Kappa analysis is required.

39. Kappa value : is a useful reliability analysis index for measuring categorical variables, which indicates the consistency of different observers' judgments on the same batch of results and the same observers' judgments on the same batch of results under different circumstances. Kappa value is different from the observed agreement rate, which takes into account the influence of chance factors on experimental consistency when judging the consisteny of two measurements.

40. The Kappa value ranges from -1 to +1. If K is negative, it means that the agreement rate of observation is lower than that caused by chance. K=-1 , indicating that the two results are completely inconsistent. K=0 , indicating that the observation agreement rate is entirely due to chance. If K > 0, it means that the consistency rate of observation is higher than that of chance. K=1 , indicating that the two results are identical. The higher the Kappa value is, the better the consistency is. The consistency below 0.2 is negligible, the poor consistency between 0.2 and 0.4, the moderate consistency between 0.4 and 0.6, the good consistency between 0.6 and 0.8, and the excellent consistency above 0.8.

41. When evaluating tests, it should be noted that a diagnostic test with better authenticity does not necessarily have better reliability, and better reliability does not necessarily have better authenticity. Authenticity and reliability are not necessarily related. Therefore, when selecting the evaluation test, we should consider its authenticity and cannot ignore its reliability.

42. （三）Evaluate the benefits of the experiment predictive value An indicator that reflects the extent to which screening results are used to estimate the Likelihood of illness and absence of client . P ositive predictive value (PPV) N egative predictive value (NPV)

43. PPV Refers to the proportion of positive test results that are actually diseased (true positive) NPV Refers to the likelihood that a negative screening test will not develop the target disease

44. Likelihood ratio It refers to the ratio of the probability of occurrence of a test result in the case group to the probability of occurrence of the test result in the non-case group. Because the test results have Yin and Yang, the likelihood ratio also has a positive likelihood ratio and a negative likelihood ratio.

45. Positive likelihood ratio ， LR+ ： It refers to the ratio of the true positive rate to the false positive rate of the test result, indicating how many times the probability that a certain test result is positive in a patient than in a non-patient. The higher the value , the greater the probability that the positive test result is true positive. For a test of high diagnostic value, LR+ should be significantly higher than 1.

46. Negative likelihood ratio ， LR- ： It refers to the ratio of the false negative rate to the true negative rate of the test result, indicating how many times the probability of a negative test result in a patient is than that of a non-patient. The smaller the value , the more likely it is that a negative test result is true negative.

47. If the LR+ value is large and the LR- value is small, the test is ideal. The ratio of LR+ to LR- can be calculated to evaluate the positive and negative of the test.

48. Due to the different levels of hospital diagnosis and treatment, and the prevalence of diseases in different regions are also different, so the prevalence of people who go to different hospitals is also different. Therefore, the predicted values of diagnostic test evaluations performed in one hospital or region cannot be extrapolated to hospitals at other levels or in other regions. However, the calculation of likelihood ratio is not affected by the prevalence rate, but only related to sensitivity and specificity. Although it is not as intuitive as the predicted value , it is an effective method in evaluating diagnostic tests in hospitals.

49. Optimized test method Select the population with high prevalence as the subjects Application of combined tests section 3 the method of improving the test efficiency

50. Select the population with high prevalence as the subjects As mentioned earlier, when sensitivity and specificity are constant, the positive predictive value of the test increases with the prevalence. Applying the test to a population with a higher prevalence can increase the number of newly detected cases, increase the positive predictive value, reduce the cost of the test, and improve the efficiency of the trial.

51. Combined test In the implementation of screening, multiple screening tests can be used to check the same subject in order to improve the sensitivity or specificity of screening and increase the benefits of screening, which is called combined testing. Series connection： If all the screening tests are positive, they are considered positive Can improve specificity In parallel： A positive result in any of the screening tests is considered positive. Can increase sensitivity

52. Table 3 Results of OB and OA combined test screening for colorectal cancer Test result Colorectal cancer patient Non-colorectal cancer patients OB OA + - 19 3 - + 23 16 + + 27 2 - - 6 69 total 75 90

53. Table 4 Results of OB and OA combined test screening for colorectal cancer Test result Colorectal cancer patient Non-colorectal cancer patients OA + 50 1 8 - 25 72 total 75 90

54. Table 5 Results of OB and OA combined test screening for colorectal cancer Test result Colorectal cancer patient Non-colorectal cancer patients OB + 46 5 - 2 9 85 total 75 90

55. Table 6 Results of OB and OA combined test screening for colorectal cancer Test result Colorectal cancer patient Non-colorectal cancer patients parallel test + 27 2 - 48 88 total 75 90

56. Table 7 Results of OB and OA combined test screening for colorectal cancer Test result Colorectal cancer patient Non-colorectal cancer patients parallel test + 71 21 - 6 69 total 75 90

57. summary Definition of screening and diagnosis. The experiment was evaluated from three aspects: authenticity, reliability and benefit. The main indexes to evaluate the authenticity of the test are sensitivity, specificity, correctness index, etc. The factors that affect the reliability of the test include subjects, observers and laboratory conditions. Principles for determining the cut-off value of positive test results. Methods to improve test efficiency.

58. What is screening and diagnosis? Discuss the purpose and application of screening and diagnosis. How are screening or diagnostic tests evaluated? What is the principle for determining the cut-off value of a positive test result? The relationship between the predicted value and prevalence, sensitivity and specificity was described. What are the ways to improve test efficiency? Review the questions

59. 1. The purpose of the screening test is () Diagnosis of suspected patients To evaluate the specificity of the screening test To evaluate the sensitivity of the screening test Prevent missed diagnosis in the screening population A B C D 提交 Identify suspected patients of a disease or people at high risk of a disease from an apparently healthy population E 单选题 1分

60. 2.parallel test can improve the diagnostic method () Specificity Sensitivity Jorden index Consistency A B C D 提交 Diagnostic accuracy E 单选题 1分 此题未设置答案，请点击右侧设置按钮

61. 3.A screening test for breast cancer was used to examine 1000 women with biopsy-confirmed breast cancer and 1000 women with breast cancer. The results were 800 positive in the group with cancer and 100 positive in the group without cancer. Missed diagnosis rate is () 80% 20% 10% 30% A B C D 提交 90% E 单选题 1分

62. 4.A screening test for breast cancer was used to examine 1000 women with biopsy-confirmed breast cancer and 1000 women with breast cancer. The results were 800 positive in the group with cancer and 100 positive in the group without cancer. specificity is () 80% 20% 10% 30% A B C D 提交 90% E 单选题 1分

63. 5.Suitable for screening in the following description are () Diseases with unclear natural history Diseases with low prevalence Early diagnosis improves prognosis of the disease serious and incurable diseases A B C D 提交 Diseases that cause mental stress to the patient population E 单选题 1分

64. Suppose a study examined 1000 people, the prevalence of a disease was 6%, and the screening test was 90% sensitive and 80% specific . 1. The false positive among those who have been screened is ( ) A.54 B.6 C.188 D.752 E.940 2. The screened negative is ( ) A. 60 B.940 C.242 D.758 E.188 3. Who was screened positive ( ) A.242 B.60 C.188 D.752 E.940 4. Number of true patients among those screened negative ( ) A.54 B.6 C.188 D.752 E.940 5. Number of non-patients who have been screened negative ( ) A.54 B.6 C.188 D.752 E.940 exercises

65. 6 . Number of real patients among those screened positive () A.54 B.6 C.188 D.752 E.940 7 . The positive predictive value of this screening test is () A.22.3% B.77.7% C.0.8% D.99.2% E.Not enough to calculate 8 . The negative predictive value of such a screening test is () A.22.3％ B.77.7％ C.0.8% D.99.2％ E.Not enough to calculate 9 . The likelihood ratio of this screening test is () A. Positive likelihood ratio 0.125, negative likelihood ratio 4.5 B. Positive likelihood ratio 4.5, negative likelihood ratio 0.125 C. Positive likelihood ratio 8, negative likelihood ratio 0.125 D. Positive likelihood ratio 4.5, negative likelihood ratio 0.222 E. Positive likelihood ratio 4.5, negative likelihood ratio 0.125

66. screening p atient nonpatient total p ositive 54 188 242 negative 6 752 758 total 60 940 1000 Data sorting 1. Among those tested positive, who were falsely positive ( 188 ) 2. Who were screened negative ( 758 ) 3. Who were screened positive ( 242 ) 4. Number of true patients among those screened negative ( 752 ) 5. Number of non-patients who have been screened negative ( 6 ) 6. Number of genuine patients among those screened positive ( 54 ) 7. The positive predictive value of this screening test is ( 54/242=22.3% ). 8. The negative predictive value of such screening tests is ( 752/758=99.2% ) 9. The positive likelihood ratio for this screening test is (54/60) : (188/940) =4.5. The negative likelihood ratio for this screening test is (6/60) : (752/940) =0.125 .

67. 《流行病学》，詹思延主编，人卫第 8 版， 2017 年 7 月 《临床流行病学》， 詹思延主编，人卫第 2 版， 2015 年 6 月 中华流行病学杂志， http:// www.cma.org.cn 中国公共卫生杂志，http://www.zgggws.com 中华人民共和国国家统计局 ： http://www.stats.gov.cn/ 公卫卫生科学数据中心：http://www.phsciencedata.cn/ 推荐书目和网站

68. 谢谢