Identifying Value-Driven Approaches to Strengthening the VA Physician Workforce

背景：初级保健医生（PCP）短缺是一个重要的和日益严重的问题， 美国医疗保健系统。预计到2025年，美国的PCP总体短缺量将高达53，000个，部分原因是 由于人口老龄化、疾病负担增加和 住院医师对初级保健职业的兴趣下降。最近，VA的PCP短缺 由于强调改善退伍军人获得高质量和及时的服务， 健康护理提供及时的初级保健需要有足够多的初级保健医生。但截至 2016年5月，据报道，VA一直试图填补多达500个PCP职位，约占12%的职位。 当前VA PCP劳动力。以前的研究没有研究几个重要的个人，经济，工作和 与卫生系统招募和留住PCP相关的系统层面因素。这些 研究不足的因素包括薪酬和补偿的组成部分，学术联系， 以病人为中心的医疗家庭模式和工作场所的气候，这代表了VA数据提供 独特的贡献。此外，没有先前的研究已经检查了因素的相对重要性， 解释PCP的招募和保留。确定和优先考虑与PCP相关的关键因素， 就业选择将有助于退伍军人事务部和其他卫生系统战略性地招聘PCP， 制定以证据为基础的战略，以保留高质量的PCP。 目的：本研究的目的是：1）识别和优先考虑个人，工作，经济和系统- 与选择VA就业和选择农村实践（即招聘）相关的水平因素 使用现有的VA数据以及为本研究收集的新的定性和调查数据，2）识别和 优先考虑与PCP长期保留相关的个人、工作、经济和系统层面因素， VA和农村VA诊所使用行政数据和现有的调查数据。 方法：这项混合方法研究将检查VA管理数据，现有VA员工调查数据， 公共卫生资源数据以及为该项目收集的新的定性和调查数据。衡量因素 与VA招募和农村实践环境的选择（目标1）相关，我们将进行结构化的 对培训期间暴露于VA环境的当前内科住院医师和新PCP进行访谈 一个永久的退伍军人职位然后，我们将开发和管理一个新的调查工具， 在定性访谈和本研究中特别感兴趣的因素中确定。然后将使用三角测量 评估定性分析中确定的关键因素和主题是否可推广到更大的 内科住院医师和新的VA PCP人群。我们将进行描述性分析， 在居民和新PCP之间的访谈和调查中确定的影响因素。多变量分析 将被用来确定与VA就业和农村考虑显著相关的因素， 内科住院医师的实践。对于目标2，我们将使用VA管理数据库来跟踪 纵向的VA PCP的就业状况。我们将确定分配给初级保健小组的所有PCP 符合最低临床全职等效（FTE）阈值并在任何时间被VA雇用的患者 在2003-2016年期间。我们将把行政数据与退伍军人事务部和公众调查数据联系起来， 利益因素。将应用生存数据分析的统计方法评估以下因素的影响： 假设因素季度保留概率和PCP的VA就业的预期持续时间。 平行分析将评估五氯苯酚在农村退伍军人设施中的存留情况。次要分析将检查 使用扩展的两部分模型的临床FTE的纵向概况。在这两个目标中，我们将应用计量经济学 包括Shapley值分解在内的方法来划分PCP补充和保留的变化 主要解释因素和因素组合。

Background: Primary care physician (PCP) shortages are a significant and growing problem confronting the US healthcare system. An overall shortage of up to 53,000 PCPs is projected in the US by 2025, due in part to increased demand for primary care services because of an aging population, growing disease burden and a declining interest in primary care careers among medical residents. A shortage of PCPs in VA has recently come under increased focus due to the emphasis on improving Veterans’ access to high quality and timely healthcare. Providing timely access to primary care requires a sufficiently large PCP workforce. However, as of May 2016, VA was reported to have been trying to fill up to 500 PCP positions, reflecting roughly 12% of the current VA PCP workforce. Prior research has not examined several important individual, economic, job and system-level factors associated with the recruitment and retention of PCPs by health systems. These understudied factors include components of pay and compensation, academic affiliation, elements of the patient-centered medical home model and workplace climate, which represent areas where VA data offer unique contributions. In addition, no prior research has examined the relative importance of factors in explaining PCP recruitment and retention. Identifying and prioritizing key factors associated with PCPs’ employment choices will help VA and other health systems with the strategic recruitment of PCPs and the development of evidence-based strategies to retain high quality PCPs. Objectives: The objectives of this study are to: 1) identify and prioritize individual, job, economic and system- level factors associated with the choice of VA for employment and selection of rural practice (i.e. recruitment) using existing VA data as well as new qualitative and survey data collected for this study and 2) identify and prioritize individual, job, economic and system-level factors associated with long-run retention of PCPs within VA and in rural VA clinics using administrative data and existing survey data. Methods: This mixed methods study will examine VA administrative data, existing VA employee survey data, public health resource data and new qualitative and survey data collected for this project. To measure factors associated with VA recruitment and choice of rural practice setting (Aim 1), we will conduct structured interviews with current internal medicine residents exposed to VA settings during their training and new PCPs with a permanent VA position. We will then develop and administer a new survey instrument using key factors identified in qualitative interviews and factors of particular interest in this study. Triangulation will then be used to assess whether key factors and themes established in qualitative analyses generalize to the larger population of internal medicine residents and new VA PCPs. We will perform descriptive analyses to compare influential factors identified in interviews and surveys between residents and new PCPs. Multivariable analyses will be used to identify factors that are significantly associated with consideration of VA employment and rural practice among internal medicine residents. For Aim 2, we will use VA administrative databases to track the employment status of VA PCPs longitudinally. We will identify all PCPs assigned to a panel of primary care patients who met a minimum clinical full-time equivalent (FTE) threshold and were employed by VA at any time during the period 2003-2016. We will link administrative data with VA and public survey data to examine key factors of interest. Statistical methods for survival data analysis will be applied to assess the influence of hypothesized factors on quarterly retention probabilities and PCPs’ expected duration of VA employment. Parallel analyses will assess PCP retention in rural VA facilities. Secondary analyses will examine the longitudinal profile of clinical FTE using an extended two-part model. In both aims, we will apply econometric methods including Shapley value decompositions to partition the variation in PCP recruitment and retention explained by key explanatory factors and factor combinations.