Study Design and Study variables
Inclusion Criteria
Qualified nurses on duty, psychiatrist registrars and psychiatrist consultants on duty at the
time of the study.
Sample Size Determination
The sample size was determined using the following formula by Fisher et al. (1935) which
assumes that the sample is randomly distributed within the population:
n=Z
2
P (1-P)
d2
Where:
n = the desired sample size if the sample size is greater than 10,000
Z= sample interval at 95% confidence limit = 1.96
p= proportion of health care workers, which is taken to be 50% since the proportion is not
known, hence p=0.5
d= the margin of error at 95% confidence limit=0.05
Substituting the values in the formulae;
n= 1.962 * 0.5 * 0.5 /0.05 = 384
Sample size before adjustment= 384
Sample size after adjustment for a population less than 10,000
Since the population of nurses and doctors working in Mathari Hospital is approximately 265,
the finite proportion correction factor was used to adjust for a population which is smaller
than 10,000 as follows:
nf = n
1+n/N
Where;
nf = the desired sample size (when the population is less than 10,000)
n = the desired sample size (when the population is more than 10,000)
N= the estimate of the population size
Therefore, nf = /nf=
385 = 157.4
1-384/265
Sample size after adjustment= 157 nurses and doctors
To accommodate for non-response, additional ten percent, which is 16 questionnaires of the
estimated sample size was added which translated to a sample size of 173 respondents.
Sampling Procedure
A sample frame of nursing staff was obtained from the Nursing Service Manager (NSM),
who is in charge of the nursing department.
To select the psychiatrist registrars and psychiatrist consultants, a serialized list of doctors
currently working in Mathari, in each of the strata was compiled using personnel records
obtained from the Human Resource (HR) office.
The study used a stratified sampling technique in selecting the study respondents to ensure
study respondents are representative of the different cadres from the sample frame. In this
sampling technique, three cadres working in the hospital were treated as strata; that is,
Nurses, and doctors, totaling 265 respondents.
The compiled list will include staff identification details such as their respective departments
for ease of tracing the respondents. Each of the serialized lists was used randomly to select
respondents from each of the strata using a table of random numbers.
Obtaining 141 nurses
Sample size from each stratum was calculated as follows:
n1 = number of respondents in stratum
n
2
= number of respondents
nf = sample size required
From a list of the total number of nurses, 221 (n1),144 (nf) nurses were selected by dividing
221 nurses by the total number of population of the three cadres, 265 (n2), then multiplying
by 173 which is the desired sample size.
Sample size = n1/n2 x nf
Systematic random sampling was further use to select every kth. Therefore:
K= n1 / n2
k= 221/144 = 1.5 ≈ 2
Every kth was selected again until the number of nurses required was attained.
Obtaining 5 Psychiatrist consultants
The total number of consultants which 8(nf), was divided by the total population of the three
cadres, then multiplied by 173, which is the desired sample size to get 5(n). Systematic
random sampling was further use to select every kth. Therefore:
K= n1 / n2
K= 8/5 = 1.6 ≈ 2
Obtaining 24 Registrars
The total population of registrars which is 34(n1) was divided by the total number of
population of the three cadres, which is 265, then multiplied by 173 which is the desired
sample size, to get 24(n2). Systematic random sampling was further use to select every kth.
Therefore:
K= n1 / n2
K= 36/24 = 1.5 ≈ 2
The proportionate sampling was undertaken in the selection of 141 nurses, 5 psychiatrists and
24 registrars. This resulted to a total of 173 that were recruited in the study (Table 3.2).
Table Sample size allocation in strata
Distribution of respondents. Calculated sample size
A. Number of Nurses = 221
221/265 x 100=83.4%
83.4/100 x 173 = 144.2
144
B. Number of Psychiatrist Consultants = 8
8/265 x 100 = 3.0%
3.02/100 x 173 = 5.2
5
C
Number or Psychiatrist Registrars = 36
36/265 x 100 = 13.6%
13.6/100 x 173 = 23.5
24
Key informants selection for the study was purposive. Key informants were selected based on
the assumption that due to their administrative position, they have experience and enough
knowledge of basic and emergency mental health care for patients. The key informants
included the facility in-charge, outpatient departmental head, male in-patient departmental
head, female inpatient departmental head and departmental head of maximum security unit.
Table Distribution of respondents in the Sample Size
N
o
Strata (Cadre) Total population Adjusted
Size/Percentage
Sample Actual
Size
Sample
A. Nurses 221 144 (83%) 131
B. Psychiatrists
8
5 (3%)
5
C. Registrars 36 24 (14%) 21
Total 265 173 (100%) 157
3.5.4. Consenting Procedure
Eligible respondents were informed about the study, and consent sought. After consent
explanation, those who gave consent were requested to sign the consent form after which a self-
administered questionnaire interview was done.
Research Instruments
The data collection instruments used were researcher-developed self-administered
questionnaires (appendix 5), and a key informant interview guide. Self-administered semi-
structured questionnaires were used to gather information on a demographic characteristic of
the respondents, structure factors, and process factors. (Appendix 5). Respondents spent
about 30 minutes each to respond to the questionnaire. Key informant guides (Appendix 6)
was used for interviews. Key informant interview guide containing five items asking similar
information, comprised probes on perceptions of psychiatric care quality and factors affecting
the quality of care. These included the structural, process and care outcome factors identified
using the. The key informant interview guide (Appendix 7) was used to obtain information on
the same areas from key persons such as NSM and departmental charges. Each interview
lasted between 20-30 minutes. Both tools were developed by the researcher. The researcher
used other comparable studies to guide and inform the development of the research tools.
Pre-testing the Questionnaire
Pre-testing of study instruments for validity and reliability was done at the Emergency
Department in Kenyatta National Hospital (KNH) in Nairobi County because of its proximity
to Mathari hospital. It is a public referral and teaching hospital where nurse and doctors
attend to emergency mental health patients amongst other cases. Being a new instrument, the
feedback obtained was used to amend and ensure the validity and reliability of the tools to
ensure it could collect the information required.
Data Collection, Management And Analysis
Data Collection Method
Two research assistants were trained and orientated to the requirements of the research.
Eligible respondents who were willing to participate signed an informed consent form.
(Appendix 3) Data were collected using self-administered questionnaires comprising four
sections with a total of thirty-eight items. Key informant face to face interviews was done.
Interviewer’s questions and the respondent's answers were recorded on a voice recorder.
Data Collection procedures
Recruitment Process
Recruitment of respondents was done by nurses, psychiatrists, and psychiatrist registrars
working in the hospital, and met the inclusion criteria. Permission was being sought from the
facility management board (after getting ethical review approval) to conduct the study. After
obtaining management approval, a list of all the health care professionals was obtained from
the human resource registry, was serialized and used to randomly sample respondents for the
study. Each recruited respondent was approached individually and informed about his/her
recruitment to participate in the study
Key informants recruited purposively based on the assumption that being in charge of the
departments they head, they have experience and knowledge in the study area. Five
interviews were done, the NSM and departmental charges from male, female, outpatient and
maximum security unit were targeted.
Selected key informants were approached individually and informed about the study, after
which a schedule list of interviewees was prepared to indicate the time and venue of the
interview.
Interview Procedure/ Data collection
Data collection (interviews) was done in a manner that respects respondents privacy by
ensuring venues of interviews were neutral such as unoccupied rooms and offices and also
ensuring that the respondent felt comfortable during data collection. Their confidentiality was
assured. They were reminded before the interview that their contribution was voluntary and
they have had an option to pull out at any point during data collection. Each questionnaire
interview lasted approximately between 15-20 minutes.
For the key informant interviews, participants who gave informed consent were interviewed
face to face using a pre-tested KI guide (Appendix 6) on a voice recorder to ensure optimal
archiving and retrieval of data. The interview schedule list was prepared and used to guide
the interviews to minimize inconveniences to the participants. Permission to use the recorder
was requested from the participant in whom confidentiality of their information was assured.
To make certain protection and privacy of the participant’s interview responses and identity,
participants were made aware that all recordings and data will be placed in a locker in an
individual designated cabinet, data stored in a password protected computer and the tapes
destroyed immediately after the research is concluded. They were informed the tapes will
only be listened to by the supervisor and researcher and will be accessible to the partaker if
they do feel to listen to their comments.
Each interview session lasted approximated to last for 20-30 minutes. After each session, the
interviewee was thanked for their time.
Data Management Analysis and Presentation
Completeness of questionnaires was checked after the interview, and clarification sought
from the respondents immediately. Quantitative data were entered into SPSS spreadsheet and
analyzed to capture the specific objectives of the study. Quantitative data analysis was done
using both descriptive and inferential statistics. The descriptive statistics were presented
using frequency tables and graphs while inferential statistics were presented using a chi-
square. Qualitative data were grouped into themes and sub-themes and analyzed. SPSS
version 22was used to enterer coded data, cleaned and analyzed based on the study
objectives’
Qualitative data were thematically analyzed manually based on the study objectives. Data
collected from key informant interviews were coded using sub-themes developed established
on the research objectives.
Ethical Considerations
Ethical approval was sought from the UON/KNH research and the ethical committee for
clearance and approval. Institutional authorization was obtained from the medical
superintendent of Mathari National Teaching and Referral hospital before data collection.
Questionnaires were administered after consent was sought. Participation in the study was
voluntary, and respondents had the right to withdraw at any stage. Anonymity was observed.
The study was assessing determinants of perceived quality of psychiatric emergency care in
Mathari National Teaching and Referral Hospital which is public health facility offering
mental health services. It did not cover other public health facilities, FBOs, private and NGO
facilities offering psychiatric emergency care. The researcher had no control over responses.
Informed consent was sought from all the recruited respondents. A consent form sheet
(Appendix 3) which contained comprehensive details about the study including purpose,
benefits, risks, assurance of confidentiality and participant’s rights was used to obtain
informed consent. Participation was voluntary and right of the participant to withdraw at any
stage of the study was held. To make certain protection and privacy of the partaker’s
interview responses and identity (confidentiality), partakers were made aware that all data
and the recordings would be placed in an individual designated cabinet (safe), data stored in a
password protected computer and the tapes destroyed after the completion of the
research. Only the principal investigator will have access to the protected data,
materials and records. The tapes will only listen to by the supervisor and researcher but
will be accessible to the partaker if they would like to listen to their comments. In
addition, no names will be used to identify the participants in the tools and final reports
hence ensuring anonymity of responses.
Dissemination Plan
The study findings has been presented at the UoN School of Nursing during the thesis
defense. A copy of the report will be furnished to the KNH/UoN Ethics Review
Committee. The findings of the research has been published in an internationally
recognized Nursing research journal hence increasing access and use to inform
relevant policies, interventions, and activities. The results have been shared and
disseminated to the hospital managers to provide the value of the new knowledge by
informing the targeted quality of mental health emergeency care improvement initiatives
in the facility.
Assumptions of the Study
The study was be based on the assumption that:
The respondents will have a positive approach towards this study which will motivate
them to provide maximum support and cooperation;
The data collection instruments were adequately valid and reliable to be able to elicit
honest and factual/unbiased responses from study respondents and
The study findings will be used by the government, insurance providers and other
stakeholders to inform appropriate policy and programmatic interventions
