Does increasing nurse-patient ratio increase in hospital patient mortality - Research Paper Example

? Does Increasing Nurse-patient Ratio Increase Hospital Patient Mortality? Problem ment The ‘Institute of Medicine’s (IOM)’ published account, “keeping patients safe: Transforming the work environment of Nurses” shows that understaffed hospitals; high nurse-patient ratio is a major factor leading to 98,000 cases of preventable patient deaths, annually, as recorded from hospitals in the United States. Physicians and doctors have also pointed out, that the cases of inadequate nurse staffs are a major hindrance to the delivery of high quality health care services. As a result of the aging nursing labor force, issues with staff retention and the difficulties experienced during the recruitment of young staffs into the field – is made worse by the present severe nursing shortage. The issue of high nurse-patient rations will continue to affect the delivery of healthcare services across the U.S (Institute of Medicine, 2004). Additionally, the nursing shortage in the USA is expected to grow to 260,000 registered nursing staff by the year 2025. A shortage of this degree will be double that experienced in the past, in the US since the 1960s. Relating the case of nurse shortage in the U.S, one of the developed nations of the world, the case has not been any different for less developed nations like Taiwan. In Taiwan, a high number of hospitals are facing the problem of low nurse-staffing. The case has been very acute in recent years, leading to the closure of a number of wards at given hospitals, according to the statistics drawn from a study by the National union of Nurses’ association (Central News Agency, 2011). Research studies on hospitals and healthcare centers show that institutions with inadequate nursing personnel show poorer patient outcomes, which compromises the care administered to patients. Decreased nurse staffing was linked to patient safety concerns and the increased incidence of adverse situations like the development of pneumonia, pressure ulcers, urinary tract infections (UTIs) (Lichtig, Knauf & Milholland, 1999), patient mortality (Aiken et al., 2002), medication errors (McGills-Hall, Doran & Pink, 2004), unplanned extubation cases (Marcin et al., 2005) and longer length of stay on average (Pronovost et al., 1999). Blegen and Vaughn (1998) discussed that, with a higher proportion of the service hours offered by registered nurses (RNs), there was a resultant decrease in the average length of stay (ALOS), cases of upper gastrointestinal bleeding, UTIs, medication errors and patient complaints, pressure ulcers, and pneumonia and cardiac arrests or shock with failure to rescue. Towards addressing the nurse shortage, the government has adopted a potential payment model to address the increasing medical expenditures. This has been affected through the imposition of measures like, the restructure of the operations of the hospitals facing financial pressure. This encompasses the adoption of strategies like the hiring of unlicensed assistant nurses or casual nurses, who are paid lower amounts as compared to the RNs, and can fill the service gaps (Sun et al., 2005). Nevertheless, it is challenging to improve nursing staff levels, even after the replacement of registered nurses with the less costly personnel. The factors leading to the unsolvable nature of the situation include the gap between the numbers of available vacancies and the number of qualified registered nurses, willing to fill these vacancies, the working conditions, the nurse burnout resulting from the shortage and low job satisfaction levels (Yin & Yang, 2002). Another issue heightening the nursing shortage is the abandonment of the nursing profession due to increased involvement in extra-patient care services like nursing informatics and utilization management. This is evident from a recent study, which showed an average vacancy level of 28% in Taiwan (Sun et al., 2005). In Taiwan, there is also legislation requiring a minimum nurse-to-patient ratio of 1:8 at hospitals. However, this minimum ratio was not affected until 2009 and is only valid during the daytime shift. As a result, hospitals may have adequate nurse-patient ratios during the daytime shift, but fall short of it during the night: the case goes to 16 and 22 patients for one nurse during the second and the third shifts, respectively. This clearly shows that the problem of high nurse-patient ratios experienced are partly due to poor management and not the in-availability of the already limited number of nursing personnel. Therefore, the study will be of great use in exposing the inefficiencies leading to the high nurse-patient ratio, which is a cause for decreased hospital outcomes and service levels, which lead to increasing mortality levels. Addressing this issue will improve the quality of services delivered to patients. The decision to revise these areas of inefficiency lies with hospital authorities, and the government, which is largely responsible for the funding of healthcare services provision. Possible solutions to the problems and inefficiencies include the revision of the legislation guiding the provision of healthcare services, for instance the mandatory nurse-patient ratio of 1:8 should be maintained at all times. Also, the revision of funding and employment models can be altered to offer more resources and more staff for the sector. An example here is the commissioning of qualified, non-registered nurses by the government and the increment of healthcare funding to be channeled into the training of nurses. The study aims at answering the question, ‘what are the implications or effects, on patient mortality, caused by acute care hospital nurse staffing? It is hoped that the inferences drawn from this study will enable policy makers arrive at more effective and informed decisions, in the area of staffing models and level variations. Through this, they will be able to maximize the quality of the care offered to patients, as well as improve nurse satisfaction levels (Yia-Wun et al., 2012). Statement of purpose This study will explore the field of healthcare service provision, towards answering the question, as to whether an increasing nurse-patient ratio results in increasing hospital patient mortality rates. The inferences drawn will be necessary for different policy makers at the hospital as well as the state level, to address the adverse effects of the increased nurse-patient ratio. Research questions 1) Does an increasing nurse-patient ratio result in an increase in patient mortality rates? 2) Can the inferences drawn on the relationship between the increasing nurse-patient ratios help healthcare policy makers? 3) Can the inferences drawn be helpful towards improving the quality of the healthcare services offered at Taiwanese hospitals? Literature Review According to Needleman et al. (2001) hospital nurse staffing is an issue of critical importance, as it is an area that affects the quality of service and the safety of patients at hospitals. Nursing-sensitive results are a major indicator on the safety of patients and the quality of care administered at a healthcare center, thus, can be defined as highly responsive to changes in nursing services administration and interventions. Some of the undesirable patient outcomes that result from ineffective nursing services, and which are sensitive to nursing care include pneumonia, urinary tract infections, shock, lengthened hospital or ICU stay, upper gastrointestinal bleeding, 30-day mortality and the failure to rescue. The failure to rescue is the incidence of the death of a patient diagnosed with a serious condition – for which early identification by nursing staff, thus the administration of the required nursing and medical interventions may have influenced the probability of death. Hickam et al. (2003) points out that a wide range of studies on the subject has established a link between high nurse-patient ratios to the higher incidence of adverse patient outcomes including death. Hickam cites the evidence from a report entitled “the effect of healthcare working conditions on patient safety,” compiled by an AHRQ-funded evidence-based practice center (EPC). The study assessed 26 studies on the association between nurse-patient ratios and the measures of patient safety. A majority of the studies explored nurse-patient ratios and the incidences of adverse occurrences at different hospitals. The conditions explored included non-fatal adverse outcomes like nosocomial infections, patient falls and pressure ulcers, as well as adverse effects like in-hospital deaths. The researchers discovered that higher nurse-patient ratios were directly related to higher incidences of non-fatal adverse outcomes. These facts were true – both at the nursing unit level and the hospital level. With regard to fatal adverse effects, the cases of in-hospital deaths were not constantly linked to the cases of higher nurse-patient ratios; there was no consistency on the relationship between mortality levels and high nurse-patient ratios. Cho et al. (2003), Needleman et al. (2002) and Kovner & Gergen (1998) have explored low staffing levels as a cause for increased levels of adverse outcomes among patients, as reported from reports compiled by the EPC under the funding of the AHRQ. The studies explored the link between adverse patient results and high nurse-patient ratios. From the five studies, it was conclusive that low nurse staffing at hospitals was a cause for one or more negative patient outcomes. The studies also reflected some statistics on the frequency of nursing-sensitive patient outcomes on patients under hospital care. The different hospitals contacted during the study reflected different adverse effects and varying rates, which varied by the patient groups in question (surgical or medical) as well as other varying factors. For instance, the urinary tract infections (UTIs) affected between 1.9 and 6.3% of the patients under surgery and pneumonia affected between 1.2 and 2.6% of the patients under surgery. According to Needleman et al. (2001), an extensive study administered under the funding of the center for Medicare and Medicaid services, the National institute of nursing research and the National health resources and services administration. The study explored the records of five million patients under medication and 1.1 million patients under surgical practice treated at seven hundred and ninety-nine hospitals in the year 1993. From the study, the principal findings included that the hospitals with high numbers of registered nurses registered less adverse patient outcomes: pneumonia, UTIs, upper gastrointestinal bleeding, shock and longer stays at the hospital, as compared to those at the hospitals with less registered nurses. The study also showed that major surgery patients at hospitals with a high number of registered nurses suffered from fewer cases of two adverse patient outcomes: failure to rescue and UTIs, as compared to those under hospitals with less RNs. The inferences also indicated that hospitals with higher numbers of registered nurses were linked to 3 to 12 percent decrease in the incidences of adverse outcomes, depending on the outcome in question, as compared to those with fewer RNs. The study also showed that higher levels of staffing at all levels of nursing was linked to a 2-25% decrease in the incidences of adverse outcomes, depending on the outcome, as opposed to the centers with less staffing. From the study, there was a direct correlation between higher nurse-patient ratios and the incidences of pneumonia, as the addition of half an hour of registered nurse staffing for a patient day could decrease pneumonia cases among surgical patients by more than 4 percent (Kovner, Mezey, and Harrington, 2000). Aiken et al. (2002) points out that despite the fact that many researches do not give consistent conclusions on the relationship between higher nursing workload and increased patient mortality, two recent studies funded by the AHRQ uncovered that the likelihood of failure to rescue and a 30-day mortality are higher at hospitals with lower nurse staffing. From the first study, it was concluded that the addition of a surgical patient into a nurse’s coverage came with a 7% increment in the likelihood of dying within a period of 30 days after admission and a 7% increment in the likelihood of failure to rescue. From the 168 hospitals covered by the study, all had a nurse-patient ratio of between 1:4 and 1:8 a total of 4,535 patients died within 30 days after admission. From the hospitals with a nurse-patient ratio of 1: 4, approximately 4,000 patients would be reported dead, while those with a nurse-patient ratio of 1:8 would report 5,000 deaths. The second study uncovered that the 30-day mortality rates reported among patients suffering from AIDS were lower at hospitals with higher nurse-patient ratios and AIDS specialty services, as opposed to those with lower nurse-patient ratios and care services. The study uncovered that a 0.25 increment in nurse per patient day resulted in a 20% reduction in the 30-day mortality rate (Aiken et al., 1999). The study uncovered that nursing staffing may be gauged by the number and proportion of registered nurses, as well as the educational level of the staffs assigned to hospitals. The study found that a 10% increment in the proportion of the nurses with an education level of bachelors’ degree was linked to a 5% reduction in the failure to rescue and the likelihood of dying within 30 days after admission, among patients under surgery (Aiken et al., 2003). Conceptual framework This study will be guided by the theoretical model developed by the “Agency for National health Research and quality,” proposing that healthcare centers and hospitals with low nursing staffs tend to present higher incidences of poor patient outcomes including the following: shock, pneumonia, urinary tract infections, cardiac arrest and in other cases higher patient mortality. From the review of literature statistics of research on the field was discussed in details, showing the association between low nurse staffing and adverse patient outcomes, including increased mortality. The model will also be guided by the nursing theory founded by Florence Nightingale, declaring that nurses should make use of all the available resources and the hospital environment to maximum, towards administering patients with quality healthcare services. The subjects for the study were drawn from medical, intensive care units, surgical wards, and comprehensive wards, sampled from 421 western-medicine hospitals based in Taiwan. The study focused on accredited western medicine-hospitals only. Data was collected from a total of 108 wards at 32 hospitals over seven month duration. The subjects will be surveyed using a questionnaire, from which hospital characteristics, ward staffing, and patient outcomes would be reported. The study seeks to explore the relationship between low nurse staffing and the incidence of adverse patient outcomes – fatal and non-fatal. The working hypothesis for the study is that, “higher nurse-patient ratios result in increased mortality rates and non-fatal patient outcomes. Aim of Study and Hypothesis The hypotheses of the study include the following: 1) An increasing nurse-patient ratio results in increasing patient mortality rates 2) The inferences drawn on the relationship between the increasing nurse-patient ratios will be helpful to healthcare policy making 3) The inferences drawn will be helpful towards improving the quality of the healthcare services offered at Taiwanese hospitals The aims of the study include uncovering the association between low nurse staffing and patient mortality rates at Taiwanese hospitals. The study also aims at improving the practice of healthcare policy making and the improvement of the services offered at these hospitals. Method of study and sample selection The data used for the current study was collected from the 2008 statistics from the “hospital nurse staffing and patient outcomes project,” funded by the Taiwan department of health (DOH). The sample for the study was drawn from different hospitals and ward types, including medical, ICU, surgical, and comprehensive wards. The study excluded psychiatric centers and hospitals which are not accredited. Through stratified random sample selection, 69 hospitals were chosen, from which the required 32 hospitals were drawn, because the anticipated participation level was expected to be low. The functional sample size after exclusion was 32 hospitals, which saw a 46.4% feedback rate. Data was collected from a total of 108 wards at the 32 hospitals, throughout a study period of seven months. The hospitals taking part in DOH projects were not willing to offer patient-level statistics. As a result, the study used wards as the unit of analysis. Monthly statistics were gathered from the study centers starting 1st July 2008 to January 31st of 2009. From the exploration we collected a total of 756 observations. Study settings The settings of the study include 108 wards of different types medical, ICU, surgical, and comprehensive wards within the 32 hospitals chosen for the study, out of the total 421 accredited western-medicine hospitals selected through the exclusion method. During the exclusion, non-accredited and psychiatric hospitals were not given consideration. Of the 32 hospitals explored during the study, those participating in DOH programs declined giving patient-level data, therefore, the study relied on ward-unit analysis. Study Design (Quantitative, non-experimental design) The conceptual model of the study placed focus on the traditional association existing between the nursing hours spent at hospitals and levels of patient mortality, which may be linked to both provider and patient characteristics. Further, the study encompasses the hospital organization and patient factors, which may influence the relationship between nurse-patient ratios and patient outcomes. All the variables chosen for the analyses were not dictated by the statistical model chosen, but were chosen from previous studies – both empirical and theoretical – in dictating the association between the examined factors and patient outcomes. The variables used for the study were introduced as follows: given that any hospital ward under study had many patients over one-month duration, the dependent variable is the binary patients’ outcome Y. Here, if any death is registered within the one-month period, the variable Y is set as one, and in the case that no death was registered, the variable Y remains at zero (0) (Hsu et al., 1992). The death count featured both 30-day mortality and in-hospital mortality. This was the case, because; estimating the mortality rate based on in-hospital deaths would most likely, be influenced by hospital discharge levels, thus could reflect lower rates of deaths, which do not reflect the effectiveness and the quality of hospital care. Out-of-hospital deaths information was gathered from linking death records and exclusive patient identifiers. The independent variables included labor variations, for example direct nursing hours, and levels of man-power; capital variables, like hospital scale; and controlled variables, like time and the severity of diseases. Direct nursing service hours represent the average nursing time that contribute to direct patient care to the patients under study, per day throughout a one-month duration, and were gathered using the computation model (Hsu et al., 1992): DNCH (direct nursing care hours) = Working hours X number of nurses Beds X occupancy rate Research design features The research design features that qualify the design chosen include the following (Hsu et al., 1992): a. The design will offer a chain of guidelines to keep the researchers in the right direction, towards securing the required association between low nurse staffing levels and increased patient mortality. b. The design focuses on important processes, therefore reducing the cases of time and resource wastage on processes that do not contribute towards the development of the study. c. The design will foster effective organization and co-ordination, as it describes the different variables and models of testing them. d. The design is flexible – to allow for modification where there is need; incorporating new relationships that may surface as, the new relationships and insights are likely to lead to the desired results. e. The design is created in a way that promotes data sourcing and the cooperation of subjects and informants. f. Most importantly, the design is fit for the study, as it is manageable for the current area of study. Intervention Based on the design features, intervention was affected to direct the study, towards the capture of the required information. An example here is the assigning of two dummy variables to distinguish the varied manpower resources. The first was using median (DNCH) direct nursing care hours where the participants were split into two different classes: the short DNCH group and the long DNCH group. The second dummy variable was described using the median count of nurses staffing at the different hospitals, differentiating the participants into two different groups: low nurse manpower classification and the high nurse manpower classification. Control over extraneous variables Control over extraneous variables was affected in the areas of the different types of wards, the time variable, the ALOS registered at the observed ward and the average age of patients. In this case, ALOS and age are used as proxies for the severity of illness. The median age was a variable used in distinguishing the subjects of the study into two groups: the younger and the older group. The median of ALOS was checked to separate the study subjects into two different classes: short ALOS group and the long ALOS group. Four distinctive types of inpatient units were used at controlling the severity of illness: these include internal medicine ward, comprehensive ward, surgical ward, and the ICU ward. In this case, a patient’s illness was considered more intensive, among ICU patients as compared to those from other groups. As a result the ICU was taken as a reference group indicating different complexity of illness from the study in-patient unit. Lastly, the time variable was affected through describing months, since the start of the study, coding them as an increasing figure, which was used to control longitudinal trends and effects over time. Statistical and Data Analysis The mixed effect logit model (MEL) The researchers used a mixed effect logit model to examine the association between patient mortality rates and nurse staffing (for instance nurse manpower and direct nursing care duration) as well as other independent variables, including the severity of illness and hospital scale. The rationale for the choice of the mixed effect model for the statistical analyses was that the data sets were gathered in the form of panel data (for instance, the same ward was studied over a period of seven months consecutively). The data sets gathered from the same ward were correlated. It is important to note that the mixed effect logit framework is a nonlinear model, thus the independent variables are categorical. As a result, the natural exponential value of the estimated coefficient was considered (Greene, 2004). Model choice and goodness of fit The intra-class relationship criterion was used to decide whether the MEL model was applicable to the data cluster gathered. The correlation model offers an indication of the proportion of variance of the incidence of death among the different hospital wards. Cohen (1998) suggested that in the case the correlation model is higher than 5.9 percent; the effect of cluster subjects (repeat panel statistics) within the same hospital ward in evaluating the regression structure cannot be overlooked. Therefore, it follows that the MEL model should be applied for the analyses. Additionally, the Hosmer-lemeshow analysis was used at testing whether the MEL model is as effective as the saturated model. References Aiken, L., Clarke, S., and Cheung, R. et al. (2003). Education levels of hospital nurses and patient mortality. JAMA, 290 (12), 1-8. Aiken, L., Clarke, S., and Sloane, D et al. (2002). Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. JAMA, 288 (16), 1987-93. Aiken, L., Clarke, S., Sloane, D., Sochalski, J., and Silber, J. (2002). Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. J Nurs Scholarsh, 288, 1987- 1993. Aiken, L., Sloane, D., and Lake, E. et al. (1999). Organization and outcomes of inpatient AIDS care. Med Care, 37 (8), 760-72. Blegen, M., and Vaughn, T. (1998). A multisite study of nurse staffing and patient occurrences. Nurs Econ, 16, 196-203. Central News Agency. (2011). Taiwan: majority of Taiwan hospitals facing shortage of nurses: poll. Taiwan News.  Cho, S., Ketefian, S., and Barkauskas, V. et al. (2003). The effects of nurse staffing on adverse outcomes, morbidity, mortality, and medical costs. Nurs Res, 52 (2),71-9. Cohen, J. (1998). Statistical power analysis for the behavioral sciences. Mahwah, New Jersey, USA: Lawrence Erlbaum Associates.   Greene, W. (2004). Interpreting estimated parameters and measuring individual heterogeneity in random coefficient models (May). NYU Working Paper No. EC-04-08. Hickam, D., Severance, S., and Feldstein, A et al. (2003). The Effect of Health Care Working Conditions on Patient Safety. Evidence Report/Technology Assessment Number 74. (Prepared by Oregon Health & Science University). Rockville, MD: Agency for Healthcare Research and Quality. May 2003. Hsu, N., Feng R., Lin, W., and Su, H. (1992). The nursing hours and time allocation in general surgical unit. VGH Nursing, 9 (4), 408-418. Institute of Medicine. (2004). Keeping Patients Safe: Transforming the Work Environment of Nurses. Washington: The National Academies Press. Kovner, C., and Gergen, P. (1998). Nurse staffing levels and adverse outcomes following surgery in U.S. hospitals. J Nursing Scholarship, 30 (4), 315-21. Kovner, C., Mezey, M., and Harrington, C. (2000). Research priorities for staffing, case mix, and quality of care in U.S. nursing homes. J Nursing Scholarship, 32 (1), 77-80. Lichtig, L., Knauf, R., and Milholland, K. (1999). Some associations of nursing on acute care hospital outcomes. J Nurs Adm, 29, 25-33. Marcin, J., Rutan, E., Rapetti, P., Brown, J., Rahnamayi, R., and Pretzlaff, R. (2005). Nurse staffing and unplanned extubation in the pediatric intensive care unit. Pediatr Crit Care Med, 6, 254-257. McGills-Hall, L., Doran, D., and Pink, G. (2004). Nurse staffing models, nursing hours, and patient safety outcomes. J Nurs Adm, 34, 41-45. Needleman, J., Buerhaus, P., and Mattke, S et al. (2001). Nurse-staffing levels and patient outcomes in hospitals. Final report for Health Resources and Services Administration. Boston, MA: Harvard School of Public Health. Needleman, J., Buerhaus, P., and Mattke, S et al. (2002). Nurse-staffing levels and the quality of care in hospitals. N Engl J Med, 346 (22), 1715-22. Needleman, J., Buerhaus, P., and Mattke, S. et al. (2001). Nurse-staffing levels and patient outcomes in hospitals. Final report for Health Resources and Services Administration. Boston, MA: Harvard School of Public Health. Pronovost, P., Jenckes, M., Dorman, T., Garrett, E., Breslow, M., Rosenfeld, B., Lipsett, P., and Bass, E. (1999). Organizational characteristics of intensive care units related to outcomes of abdominal aortic surgery. JAMA, 281, 1310-1317. Sun, C., Lin, P., Kao, C., Change, T., and Shaw, F. (2005). Practice survey of nurses in Taiwan. Department of Health, Executive Yuan (Report No. DOH94-NH-36). Taipei: Department of Health, Executive Yuan. Yia-Wun, L., Wen-Yi, C., Jwo-Leun, L., and Li-Chi, H. (2012). Nurse staffing, direct nursing care hours and patient mortality in Taiwan: the longitudinal analysis of hospital nurse staffing and patient outcome study. BMC Health Services Research, 12, 44. Yin, J., and Yang, K. (2002). Nursing turnover in Taiwan: a meta-analysis of related factors. Int J Nurs Stud, 39, 573-581. Read More