Improving Health Quality
The main research question will be restated as follows:
Is there a relationship between the quality of healthcare to ICU patients and the costs?
The quality of healthcare is often associated with an increase in the cost of healthcare because providing better quality care requires more resources of higher quality. Therefore, the null as well as alternate hypothesis will be as follows:
Null Hypothesis
There is no relationship between the quality of healthcare to ICU patients and the costs.
H0: β = 0
Null Hypothesis
There is a relationship between the quality of healthcare to ICU patients and the costs.
H0: β ≠ 0
Independent and Dependent Variables
This study seeks to assess whether cost reduction is associated with a compromise of the quality of healthcare. Therefore, the dependent variable is the healthcare cost while the independent variable will be the quality of healthcare.
Extraneous Variables
This study is only interested in the relationship between the cost of healthcare and the quality of healthcare services. However, it acknowledges that there are other variables that are likely to influence the cost of healthcare other than the quality of healthcare. The identified variable includes; the length of time a patient is hospitalized. The longer a patient is hospitalized, the higher will be the healthcare costs because it directly influences the medical supplies, nursing care hours and provision of personal effects and food. Therefore, there is a increasing relationship between the time of hospitalization and healthcare costs. Age will also influence the healthcare costs. Older people have weaker systems that will require much more support and nursing care thus increasing healthcare costs. Gender is also expected to influence healthcare cost. Samadi & Enayatollah (2013) that there is an increasing relationship between age and healthcare.
Chaudry, B., et al., (2006) reveal that there is a relationship between ICT and efficiency, quality, and cost of healthcare that is effected through various channels. The various hospitals have different healthcare systems and resources. Therefore, hospital differences are likely to influence the cost of healthcare. The study will control for these extraneous variables by using them as control variables. The list of extraneous variables maybe updated if other cofounding variables are identified from the literature.
Data Cleaning and Missing Data
The respondents will be given questionnaires to answer on their own. Therefore, those patients who have already recovered will be used as the sample. It is likely that some questions may not be filled which will result in missing data. The respondents will be called using the telephone number that was indicated on the form and interviewed about the questions. Resampling will be conducted to replace respondents who will be unreachable.
The data will be cleaned to remove outliers from the dataset. Unreasonable responses will also be eliminated from the dataset. Random resampling will be used to replace the study participants’ data that has been removed.
Data Analysis
SPSS will be used to conduct data analysis. Data analysis will be conducted in two stages. The first stage will use descriptive statistics to create an overview of the demographic profile of the selected sample. In the second stage, inferential statistics will be applied to answer the research question.
Descriptive Statistics
The questionnaire will include questions asking clients about their demographic profile. The demographic information that will be collected includes income level, gender, education level, and age. Participant ages will indicate their age in years. Education will be measured as years of formal schooling. Income level will be measured as dollars per annum. Gender will be measured as a dummy variable with 0 representing females and 1 representing males. Central tendency measures and dispersion measures will be used to explain the demographic characteristics with the exception of gender. Gender will be analyzed using frequency distribution and percentages.
Data Analysis for Study Variables
The main study variables are the cost of provision of healthcare to ICU patients and the quality of the resultant healthcare. A multi-regression analysis will analyze the data. A t-test for regression coefficients will be used to answer the key study question. The cost of healthcare will be measured in US dollars. The data will be obtained from the initial community hospitals billing department. Information on the costs billed to each of the sampled respondents will be recorded. The quality of healthcare will be measured using a 10-point Likert Scale. Respondents will be given a statement to the effect “I received high-quality service”. They will then be required to select a multiple choice that correspondents to the extent of agreement or disagreement with the statement.
A multiple linear regression will then be conducted that includes the identified extraneous variables as control variables. The obtained co-efficient for the quality of healthcare will then be tested at 5% significance level. If it is significant, then there is a relationship between the cost of healthcare provision to ICU patients and the quality of healthcare. The sign of the coefficient will allow us to assess the nature of the relationship (positive or negative).
Assumptions and Mitigation
Inferential statistics will be applied on the study sample to make conclusions about the population. Therefore, the sample is assumed to represent the entire population. A random sample strategy will be used to obtain the study participants. The strategy is appropriate will ensure that the sample is representative (Emerson, 2015). Inferential statistics also assumes that the population is normally distributed. This study assumes that population is normally distributed since it is large.
The study will use a multiple linear regression. Therefore, the standard OLS assumptions will apply. The study assumes that the variance of the errors is constant (there is no heteroscedasticity). In other words, the variance of errors does not change as the intendent variable increases. A plot of residuals versus fitted will be used to test this assumption. The assumption will be met if the residuals data points do not form a unique pattern. If they form a unique pattern, then the assumption has not been met. The other assumption is that errors are normally distributed and have a mean of zero. A Q-Q plot will be used to assess this assumption. The assumption will be met if most of the data points are along the 45-degree line. If they steer away from the line, then the assumption will be violated.
If the assumptions are violated, then an OLS specified model is not the best fit for the data. Therefore, an alternative model specification that best suits the data will be applied. AIC will be used to model specification that provides the most information given the dataset that is available. The other model specifications that will be tested include quadratic, log-log and log models.
References
Chaudry, B., et al, (2006), Systematic review: impact of health information technology on
quality, efficiency, and costs of medical care. Annals of Internal Medicine, 144(10), 742-752.
Emerson, R. (2015). Convenience Sampling, Random Sampling, and Snowball Sampling: How Does Sampling Affect the Validity of Research? Journal of Visual Impairment & Blindness, 164-168.
Samadi , A., & Enayatollah , H. (2013). Determinants of Healthcare Expenditure in Economic Cooperation Organization (ECO) Countries: Evidence from Panel Cointegration Tests. International Journal of Health Policy Managment, 63-68.
