叶鹏鹏, 汪媛, 金叶, 段蕾蕾. 不同方法在基于医院门/急诊伤害监测病例漏报估计中的比较研究[J]. 疾病监测, 2018, 33(3): 254-259. DOI: 10.3784/j.issn.1003-9961.2018.03.020
引用本文: 叶鹏鹏, 汪媛, 金叶, 段蕾蕾. 不同方法在基于医院门/急诊伤害监测病例漏报估计中的比较研究[J]. 疾病监测, 2018, 33(3): 254-259. DOI: 10.3784/j.issn.1003-9961.2018.03.020
Ye Pengpeng, Wang Yuan, Jin Ye, Duan Leilei. Comparison of different methods in estimating under reporting of injury cases in injury surveillance based on outpatient and emergency departments[J]. Disease Surveillance, 2018, 33(3): 254-259. DOI: 10.3784/j.issn.1003-9961.2018.03.020
Citation: Ye Pengpeng, Wang Yuan, Jin Ye, Duan Leilei. Comparison of different methods in estimating under reporting of injury cases in injury surveillance based on outpatient and emergency departments[J]. Disease Surveillance, 2018, 33(3): 254-259. DOI: 10.3784/j.issn.1003-9961.2018.03.020

不同方法在基于医院门/急诊伤害监测病例漏报估计中的比较研究

Comparison of different methods in estimating under reporting of injury cases in injury surveillance based on outpatient and emergency departments

  • 摘要: 目的 采用不同方法估计基于医院门/急诊伤害监测病例漏报率,比较不同方法估计结果的差异并分析原因,以期对相关研究提供线索和思路。方法 采用2次独立简单随机抽样的捕获再捕获(方法1)、1次独立简单随机抽样的捕获再捕获(方法2)、样本间接估计(方法3)和样本直接估计(方法4)4种方法,通过数据模拟,对符合一定假设情形下的医院门/急诊伤害病例漏报率进行估计,并比较不同方法点估计和区间估计结果的差异。结果 方法1漏报率的点估计和区间估计结果最差,且工作量较大;方法2漏报率的点估计和区间估计结果优于方法1,且工作量也低于方法1;方法3漏报率的点估计和区间估计结果优于方法1,但劣于方法2,工作量最小;方法4漏报率的点估计和区间估计结果与方法2接近,均优于方法3,工作量略低于方法2。结论 捕获再捕获方法对门/急诊病例中伤害病例占比有要求,提高每次抽样选取的病例数量,尤其是第2次抽样的病例数量可以提高漏报率估计精度。用样本直接估计总体漏报率比先估计总量再估计率,减少了抽样误差,可以提高估计精度。每种方法增加样本量均可以提高估计精度,但会增加工作负担。根据结果,建议采用方法4进行医院门/急诊伤害监测病例漏报估计。

     

    Abstract: Objective To understand the effects of different methods in estimating the underreporting rate of injury cases in injury surveillance based on outpatient and emergency departments and influencing factors,and provide evidence for the related research in future. Methods Four methods,capture-recapture with two independent simple random sampling, capture-recapture with one independent simple random sampling,sample based indirect estimation and sample based direct estimation,were used to estimate the underreporting rate of injury cases in hospitals under certain assumptions by conducting data simulation. The differences in point and interval estimation results of the 4 methods were compared. Results The first method had the worst point and interval estimation results and largest workload. The second method had better point and interval estimation results and less workload compared with the first method. The point and interval estimation results of the third method were better than those of the first method but worse than those of the second method,but the workload was least among four methods. The point and interval estimation results of the fourth method were very close to the results from the second method and it had less workload compare with the second method. Conclusion The use of capture-recapture methods had the requirement on the proportion of injury cases in outpatient and emergency departments. Increasing the number of outpatient and emergency cases from each random sampling,especially in the second random sampling,could enhance the precision of the estimated underreporting rate. Using sample based estimation,the introduction of sampling error could be reduced and the estimation precision could be improved compared with the estimation of total number of injury cases firstly. For each method,increasing the sample size could improve the precision of the estimated underreporting rate but add more workload. Based on the results,the fourth method is recommended to estimate the underreporting rate of injury cases in outpatient and emergency departments.

     

/

返回文章
返回