Interprofessional Care Team, Staffing, and Setting Characteristics that Impact Patient Outcomes: A Review of Reviews

Step 1. Establish the research questions

Using our team’s expertise and directives from the legislature, we narrowed our research questions to the following:

1. How were team members from different professions and roles (clinical and nonclinical) represented in studies assessing the impact of care teams on patient outcomes?

2. What community, hospital/organizational, care team, and patient factors were included in description or analysis of staffing as it impacted patient outcomes, and how were these variables defined?

3. What analytic strategies were used to examine the impact of teams or staffing on patient outcomes?

Step 2. Identify relevant studies

The search strategy was developed and executed in consultation with an experienced research librarian (CM). The original search occurred on March 3, 2022 and was most recently updated on March 30, 2023. The search strategy was created for PubMed (which includes Medline, PubMed Central, and other resources) using a combination of Medical Subject Heading (MeSH) terms, keywords, and phrases (see the complete search strategy in Supporting Information). The search terms targeted health workers, health research, hospitals/hospital settings/hospitalization, and patient outcomes. The PubMed search terms were translated for Cumulated Index to Nursing and Allied Health Literature (CINAHL) and Embase using their respective thesaurus terms and advanced search features. A manual search using the reference lists of retrieved citations was conducted for other relevant studies.

Table 1 summarizes the participants, interventions, comparison, outcomes, and study design (PICOS) used to define the inclusion and exclusion criteria strategy. Eligible studies included systematic reviews that examined interventions with a workforce component and their impact on adult acute care patient outcomes and were published within the last 10 years. We excluded all other study or publication types, such as cross-sectional studies, randomized control trials, commentaries, editorials, letters to the editor, abstract proceedings and reviews that did not include workforce personnel or patient outcomes. We restricted our search to only include reports that were published in English.

3. Select Relevant Studies

We compiled the search results in EndNote and removed duplicates, then transferred citations to Covidence software for screening. The screening process was composed of two phases: first, we reviewed the title and abstract for relevance using the inclusion criteria; and second, we reviewed the full text of remaining articles. In each of these phases, two reviewers (AP, ZG) independently screened the articles and a third reviewer (SI) assisted in resolving any discrepancies.

4. Charting the Data

Project directives informed the selection of variables to include analysis of the impact of the number, type, education, training, and experience of acute care hospital staffing personnel on patient mortality and patient outcomes. Control variables included factors such as access to equipment, patients’ underlying conditions and diagnoses, patients’ demographic information, the trauma level designation of the hospital, transfers from other hospitals, and external factors impacting hospital volumes.(12) Based on an initial review of papers to identify key components, the reviewers used a standard data extraction sheet to record the following categories: study design, data sources, inclusion criteria within that review, interprofessional team members, types of patients (e.g. acute versus intensive care), patient outcomes (e.g., length of stay, hospital acquired infection), additional factors considered in the review or analysis (e.g., equipment, technology, external factors), and analytic strategy or model. In addition, any reported challenges, limitations, strengths, or recommendations to each were extracted.

5. Synthesis of Results

We used a descriptive approach to report the data on the variables, approaches, and healthcare team composition. Given the expected heterogeneity of systematic reviews found in our search and our goal to simply identify the presence and use of variables rather than the direction of any relationships between them, a quantitative summary measure of results was not undertaken. We summarized frequency of use, definitions, and concordance of analytical strategies across the studies. We synthesized the findings into main concepts, grouped by variable categories (i.e., hospital characteristics, patient characteristics, staffing, or patient outcome). We also summarized reported challenges, limitations, strengths, and recommendations for each variable. Results informed the development of a baseline causal model that was used to guide subsequent stakeholder interviews and to develop and refine the analytic strategy for the state workforce study.

Interprofessional Teams and Staffing

The reviews provide limited evidence on the impact of interprofessional healthcare teams on patient outcomes. Most studies either focused exclusively on nursing staff (10/12) or considered direct care providers such as physicians or respiratory therapists (2/12) without mentioning the impact of indirect healthcare workers such as nutrition or environmental services. One study of interprofessional teams in the emergency department defined the healthcare team as an assortment of care professionals including occupational and physical therapists and medical social workers.(20) One review examined the relationship between nighttime intensivist staffing and patient outcomes in intensive care.(21) Two reviews included a limited number of studies examining relationships between non-registered nurse team members and quality indicators or discharge.(19, 22)

Studies that quantified the impact of nursing staff on patient outcomes used various measures to quantify the patient care workload. The most common measure of absolute staffing levels was nurse-to-patient ratio (8/12),(8, 17–19, 22–25) followed by Nursing Hours Per Patient Day (6/12).(16–18, 22, 24, 25) Other less commonly used measures for nursing workload included nurse overtime (2/12),(16, 26) Nursing Activities Score (1/12),(18) number of hours worked by nurses (1/12),(26) and Therapeutic Intervention Scoring System-to-nurse ratio (1/12).(24)

Nurse characteristics

As part of assessing staffing, reviews addressed characteristics of nursing staff and workplace factors. Assessments of nursing work environment (as measured by the Nursing Work Scale) were included in 4/12 reviews.(16, 22, 23, 25) Another common characteristic was education (4/12), or the percentage of RNs with a baccalaureate or higher degree in nursing.(16, 22, 23, 25) Three of those reviews also looked at years of nursing experience,(16, 22, 25) 2/12 included specialty certification, (22, 25) and 1/12 included tenure at the organization.(25) Job-related training or knowledge was included in 3/12 reviews,(17, 19, 25) 3/12 looked at nursing skill mix, or the percentage of nursing staff who were registered nurses,(16, 17, 22) and 1/12 included nurse reported quality measures.(25) Most studies used these elements as independent variables, while 1/12 used them as controls for their main findings.(19)

Community, Hospital, and Patient Characteristics

In general, community characteristics such as rurality or population demographics were not used as independent variables or controls when relationships between healthcare teams and patient outcomes were assessed; only 1/12 reviews included community characteristics as external factors that may affect hospital volume.(25) Within hospitals, characteristics such as equipment and technology, size, and teaching status were examined in 4/12 reviews.(16, 17, 19, 25) Two reviews controlled for equipment and technology, with definitions ranging from counts of resources such as critical care beds to facilities for major organ transplant or open-heart surgery.(16, 25) One review included hospital-level structures for discharge.(19) In all, community and hospital variables were not consistently represented or defined across the included reviews.

Patient-level characteristics were used as descriptive and control variables. 4/12 reviews considered measures related to patient health status at baseline or severity of illness during hospitalization.(17, 19, 21, 25) Only 2/12 reviews addressed other patient variables such as demographics and healthcare logistics such as insurance or finances(19, 25) and 1/12 looked at admission type (e.g., elective versus emergent admissions).(17) No included studies referenced or defined SDOH or discussed health equity.

Patient Outcomes

Patient outcomes were treated as a dependent variable in most of the included reviews (9/12), and were defined as mortality, adverse events, and other care outcomes. We used the National Quality Forum classification for adverse events, which defines 29 possible serious, reportable events that are considered largely preventable and therefore indicative of inadequate healthcare safety mechanisms.(27) Patient mortality was included in some form in 11/12 reviews. Of those, 5/12 looked at inpatient mortality,(17, 18, 21, 22, 25) 5/12 looked at post-discharge mortality at 1-month(17, 18, 20, 23, 26) or 1/12 at 1 year,(20) 1/12 did not specify a timeframe,(24) and 1/12 noted the use of mortality in many of the included studies but did not treat it as a predictor or outcome.(19) Non-fatal adverse events related to inpatient care were used in 7/12 reviews as a staffing-related outcome. Reported adverse events included healthcare associated infection (6/12),(17, 18, 22, 24–26) pressure injury (5/12),(17, 22, 24–26) patient falls (3/12),(22, 25, 26) medication errors (2/12),(22, 25) and gastrointestinal bleeds (1/12).(26) Other patient outcomes were length of stay (7/12),(17–22, 25) hospital readmission (6/12), (17, 19–22, 25) satisfaction (5/12), (18, 20, 22, 25, 26) and ventilator days (3/12).(17, 18, 21) Two studies focused on missed care(8, 16) and another included health-related quality of life as a patient outcome.(20)

Strategies for Analysis

Quantitative strategies for analysis were only addressed in 5/12 reviews. Of the reviews that extracted information on quantitative analysis, multivariate logistic regression was the most used model type in the evaluation of staffing impacts on patient outcomes (3/12),(17, 23, 26) followed by multivariate linear regression (2/12)(19, 24) and negative binomial regression (2/12).(17, 26) One of 12 reviews reported use of other statistical analyses, including the Cox proportional hazard model and hierarchical mixed effects survival.(17)

Causal Model

Study findings were used to develop an initial causal model that was used throughout the WA workforce study the Washington Acute Care CHaracteristics and patient Outcomes (WACCHO) model (Figure 2). The model presents the different categories of variables present in the current literature. We used this model to guide our stakeholder interviews and focus groups and determine if any factors were missing or inadequately represented in the model.

• Download figure
• Open in new tab

Figure 2.

Initial causal model based on review findings.

Care team composition

A key finding from this review was that interprofessional acute care team composition is undertheorized and underrepresented in current scientific literature. Because studies that included interprofessional team members occurred in specific areas outside of acute care or focused on a specific event such as discharge, we found no evidence describing the effects of a comprehensive, interprofessional care team on patient outcomes in acute care. Furthermore, studies predominantly focus on the effects of nurse staffing and infrequently consider the influence of non-nursing staff or indirect care providers such as environmental services on patient outcomes. These findings reinforce trends in current workforce analysis to focus exclusively on nursing personnel without acknowledgement of associated factors that impact the RN workload.(7) While a focus on nursing is warranted due to their essential role in acute care, research that fails to account for other factors and team members impacting RN work will miss important elements that influence the relationship between staffing and patient outcomes. For example, in the absence of other care team members, RNs often take on non-nursing roles, diluting time that could be spent on activities that optimize their full scope of practice and increasing frustration and burnout.(9) Understanding the impacts of other types of staffing is therefore essential to understanding how outcomes are affected in settings where RNs are able to exclusively focus on RN work and settings where they routinely take on other tasks. A national workplace-focused survey in the United States found that among 5,461 acute care RN respondents, 27% rated availability of appropriate ancillary staff as ‘seldom’ or ‘never’, 39% indicated availability as ‘sometimes’, and only 29% rated availability as ‘often’ or ‘always’.(28) In the context of staffing shortages and nursing burnout, organizations need to understand the optimal care team composition for both patient and worker outcomes.

Community, hospital, and patient characteristics

Nearly half of the reviews included in the study did not address community, hospital or organizational, or patient characteristics. When hospital characteristics were included in reviews, there was little consensus on how to best define and control for characteristics that may influence the provision of high-quality care. For example, hospitals equipped to perform open heart surgeries were used as proxy for more advanced equipment and technological capacity than those who do not perform open heart surgery, but it is unclear whether this denotes a substantive difference in organizational capacity or resources compared to a hospital that does not perform open heart surgeries. Data suggests that community level factors such as urbanicity, county socioeconomic status, and housing type may shape the patient population served by the organization, which impacts both care delivery and outcomes,(10) so these factors need to be considered when examining the impacts of staffing and workforce issues.

None of the included reviews addressed SDOH as patient characteristics, nor did they use the term ‘health equity’ in describing patient outcomes. This may have occurred in part because of the temporality of the reviews and the papers they sourced, but is nonetheless a gap in our understanding of relationships between staffing and patient outcomes. In general, patient characteristics such as demographics were not consistently included in staffing models, and more nuanced patient-level metrics such as insurance status or admission type were even less commonly included. Measures intended to reflect patient acuity, such as an adjustment for comorbidities or severity of illness, were similarly sparse. In the existing literature, some studies reported diagnostic codes or acuity ratings to quantify workload related to patient care, but even these measures failed to identify patients requiring extra time and personnel resources to achieve similar outcomes.(29) This implies an over-reliance on patient demographics such as age, sex, and race as proxies for SDOH and therefore limits the interpretation of results and their impact on health equity. This pattern may be related to a lack of available data in primary research studies endeavoring to quantify these relationships. In Washington State, House Bill 1272 requires hospitals to begin reporting additional patient demographic information in 2023. This data should improve future researchers’ ability to assess health equity as a critical outcome.

Variable definition and analytic strategy

Definitions of staffing metrics used by primary studies were rarely provided in the systematic reviews, suggesting that these variables were not consistently defined in workforce research. The lack of clear definitions or consistent measures makes further analysis and application of findings difficult. For example, the most common staffing metric used in the study was nurse-to-patient ratio. This ratio does not overtly reflect patient acuity as well as measurements such as Nursing Hours Per Patient Day, and therefore hinders the comparison of staffing ratios and patient outcomes among different populations and in different acute care settings.

Along with inconsistent variable definitions, we found a lack of consensus regarding the analytic strategy or model type used in studies exploring the impacts of healthcare staffing on patient outcomes. Most included reviews did not report the methods of quantitative analysis used by primary workforce studies. Of those reporting analytic strategies, multivariate logistic regression was the most common method of analysis. Future studies in this field could explore which model types and methods best reflect the relationship between healthcare staffing and patient outcomes, particularly when examining the influence of community, hospital, or patient characteristics.

Limitations

A review has inherent limitations, since findings generally rely on the accuracy and completeness of the review, and quality may vary across studies. For example, reviews may have poorly specified inclusion and exclusion criteria or an inadequate search process. This study relied on the information presented in the papers we reviewed, which may have missing information from primary studies or may be interpreted differently in our findings. Additionally, despite careful development of a systematic search strategy, we may have missed some relevant systematic reviews. Despite the various limitations, a key strength of this report is the sheer volume of primary studies assessed. More than 575 studies were represented in the included reviews. Examination of reference lists showed that 13 primary studies were represented in multiple reviews, with 9/13 in 2 reviews, 3/13 in 3 reviews, and 1/13 in 4 reviews.

However, assessment of each review showed that findings were not exclusively drawn from duplicated studies. Altogether, this review offers a comprehensive overview of current studies about interprofessional team composition and staffing and their impacts on patient outcomes in the hospital setting.