Health Data Anlaysis

image source: rand.org importing libraries

library('doBy')
library('lattice')
library('dplyr')
## 
## Attaching package: 'dplyr'
## The following object is masked from 'package:doBy':
## 
##     order_by
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union
library('tidyverse')
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ forcats   1.0.0     ✔ readr     2.1.4
## ✔ ggplot2   3.4.2     ✔ stringr   1.5.0
## ✔ lubridate 1.9.2     ✔ tibble    3.2.1
## ✔ purrr     1.0.1     ✔ tidyr     1.3.0
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter()   masks stats::filter()
## ✖ dplyr::lag()      masks stats::lag()
## ✖ dplyr::order_by() masks doBy::order_by()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(data.table)
## 
## Attaching package: 'data.table'
## 
## The following objects are masked from 'package:lubridate':
## 
##     hour, isoweek, mday, minute, month, quarter, second, wday, week,
##     yday, year
## 
## The following object is masked from 'package:purrr':
## 
##     transpose
## 
## The following objects are masked from 'package:dplyr':
## 
##     between, first, last

Loading Data Set

health_rec <- read.table('https://raw.githubusercontent.com/xrander/Health-data/master/Health_data.csv',
           header = T,
           sep = ',',
           na.strings = NA,
           dec = '.',
           strip.white = T)

health_rec <- setDT(health_rec)
head(health_rec)

##Data Explorartion

str(health_rec) # data structure

summary(health_rec) # quick statistics

nrow(health_rec)

Dealing with NA’s for outcome

Some of the data is having NA, that happens in big data, outcome is the most important here and cannot be left as NA

health_rec[is.na(health_rec$outcome),]
health_rec <- health_rec[!is.na(health_rec$outcome),] #removing the missing outcome data

length(health_rec$outcome) 
## [1] 1176

There is no missing data again in the part of the outcome. – The data have reduced to 1176 from 1177

Data transformation

some data are categorical data and they’ve been changed to factor type of data

health_rec$outcome  <-as.factor(health_rec$outcome)
health_rec$gender <- as.factor(health_rec$gender)
health_rec$COPD <- as.factor(health_rec$COPD)
health_rec$Renal_failure <-as.factor(health_rec$Renal_failure)
health_rec$hypertensive <- as.factor(health_rec$hypertensive)
health_rec$atrialfibrillation <- as.factor(health_rec$atrialfibrillation)
health_rec$CHD_with_no_MI <- as.factor(health_rec$CHD_with_no_MI)
health_rec$diabetes <- as.factor(health_rec$diabetes)
health_rec$deficiencyanemias <- as.factor(health_rec$deficiencyanemias)
health_rec$depression <- as.factor(health_rec$depression)
health_rec$Hyperlipemia <- as.factor(health_rec$Hyperlipemia)
health_rec$Renal_failure <- as.factor(health_rec$Renal_failure)

Assigning values to codes

We assigned some values to the codes to aid easy analysis

health_rec$outcome <- ifelse(health_rec$outcome == '0', 'Alive', 'Dead')
health_rec$gender <- ifelse(health_rec$gender == '1', 'Male', 'Female')
health_rec$atrialfibrillation <- ifelse(health_rec$atrialfibrillation == '0',
                                        'Having', 'Not having')
health_rec$depression <- ifelse(health_rec$depression == 0, 'Having',
                                'Not Having')
health_rec$hypertensive <- ifelse(health_rec$hypertensive == '0', 
                                  'Having','Not Having')
health_rec$Renal_failure <- ifelse(health_rec$Renal_failure == '0',
                                   'Not having','Having')
health_rec$Hyperlipemia <- ifelse(health_rec$Hyperlipemia == 0,
                                  'Having','Not Having')
health_rec$deficiencyanemias <- ifelse(health_rec$deficiencyanemia == 0,
                                       'Having', 'Not Having')
health_rec$diabetes <- ifelse(health_rec$diabetes== 0, 'Having','Not Having')

QUESTIONS

Question 1

Which age is the most in the hospital?

sort(table(health_rec$age), decreasing = T)
## 
##  89  84  81  80  78  83  85  79  82  87  75  77  86  72  66  76  88  63  64  69 
## 141  48  43  40  39  39  38  35  34  33  32  32  32  30  28  27  27  26  26  25 
##  68  67  71  65  62  70  73  74  53  56  55  60  61  58  48  57  52  51  54  59 
##  23  22  22  21  19  19  18  18  16  16  15  15  15  13  11  10   9   8   8   8 
##  47  45  46  50  90  91  92  93  96  35  37  42  43  44  38  41  19  39  40  49 
##   7   6   6   6   6   5   5   5   5   4   4   4   4   4   3   3   2   2   2   2 
##  94  97  25  28  32  95  98  99 
##   2   2   1   1   1   1   1   1
# Visuals
barplot(sort(table(health_rec$age), decreasing = T),
        col = 'blue',
        main = 'Age distribution of patients',
        ylim = c(0, 160),
        ylab = 'Number of patients',
        xlab = 'Age of Patients')

ANSWER Patients that are 89 years of age are the highest.

Question 1b

Which age group is the most in the hospital?

breaks <- seq(from = 1, to = max(health_rec$age), by = 10) #this starts 
#at the lowest age and ends at the lowest
  
age_interval <- cut(health_rec$age, breaks = breaks,
                    right = FALSE,
                    include.lowest = TRUE) # this makes the class interval from the list
age_group <- cbind(data.frame(age_interval))

age_groups <- sort(table(age_interval), decreasing = T) # generating frequency table

barchart(age_groups,
         horizontal = FALSE,
         xlab = 'Age_interval',
         ylab = 'Frequency',
         ylim = c(0, 450),
         col = 'red')

ANSWER Patients between the age 81 to 91 are the highest in the hospital with age 89 been the highest

Question 2

which age group of patients dies more in the hospital?

dead_patients <- health_rec[outcome == "Dead",c("outcome","age","gender") ]

The result shows the frequency table for dead patients according to age.

barplot(sort(table(dead_patients$age)),
        col = 'purple',
        xlab = 'age (years)',
        ylab = 'Frequency',
        main = 'Distribution of Age for Dead Patients',
        ylim = c(0, 30))

ANSWER Patients at age 89 are the highest to die, with 23 patients dying

Question 3:

What is the most prevalent gender in the hospital

gender <- as.data.frame(sort(table(health_rec$gender), decreasing = T)) # This gives a frequency of the genders
  ## Female gender is more prevalent with 618 patients

colnames(gender) <- c('gender','num') #changing column names

Question 3b:

what is the death rate of both gender

gender_dead_count <- as.data.frame(sort(table(dead_patients$gender))) #Frequency for dead patients according to gender
colnames(gender_dead_count) <- c('gender', 'count')

gender[3] <- gender_dead_count$count

names(gender)[3] <- 'dead' # this is similar to using colnames

gender$rate <- gender$dead/gender$num * 100 #rate of death per hundred

It seems the male gender is having more dead rate than female

barplot(gender$rate,
        names.arg = gender$gender,
        xlab = substitute(paste(bold('gender'))),
        ylab = substitute(paste(bold('death rate per hundred'))),
        main = 'Death rate of genders',
        col = c(5,7))

Question 4

Gender group with the highest number of death?

sort(table((dead_patients$gender)), decreasing = T)
## 
##   Male Female 
##     80     79

the gender with the highest death is male with 80 individuals dead

Question 5

How many patients died in the hospital with atrialfibrillation?

dead_patients_atrial <- as.data.frame(health_rec %>% select(outcome, atrialfibrillation) %>% 
  filter(outcome == 'Dead') %>% group_by(atrialfibrillation)%>%
  summarize(length = length(atrialfibrillation)))

dead_patients_atrial

67 patients died having atrialfibrillation

Question 5b

patients having atrialfibrillation

atrial <- as.data.frame(health_rec %>% select(gender, atrialfibrillation) %>% 
  filter(atrialfibrillation == 'Having') %>% group_by(gender)%>%
  summarize(atrial = length(atrialfibrillation))) # this gets the number of having atril... for each gender

gender <- merge(gender, atrial, by = 'gender', all = T)

645 is the number of patients having atrialfibrillation

Question 5c

Number of patients that died between the gender having atrialfibrillation

dead_patients_atrial_gender <-
as.data.frame(health_rec %>% select(outcome, gender,
atrialfibrillation) %>% filter(outcome == 'Dead',
atrialfibrillation=='Having') %>% group_by(gender) %>%
summarize(died_having_atrial = length(gender)))

dead_patients_atrial_gender

39 female and 28 male died having atrialfibrillation

gender <- merge(gender,
dead_patients_atrial_gender, by = 'gender', all = TRUE)

merging the result to gender table

Question 6

How many patients in the hospital have depression?

(sort(table(health_rec$depression)))  
## 
## Not Having     Having 
##        140       1036
barplot(sort(table(health_rec$depression)),
ylab = 'number of patients', xlab = 'depression', main = 'patients
having depression', col = 'green')

1036 out of 1176 patients are having depression

Question 6b1

show the total number of depressed male and female

depressed_individuals <- health_rec %>% select(gender, depression)%>%
  filter(depression == 'Having') %>% 
  group_by(gender) %>%
  summarize(depressed_individuals = length(depression))
depressed_individuals
gender <- merge(gender, depressed_individuals, by = 'gender', all =
TRUE) #merging result to gender table

Question 6b2

show the number of patients having depression that are dead according to gender.

dead_depressed <- as.data.frame(health_rec %>%
                                  select(outcome, gender, depression) %>%
                                  filter(outcome == 'Dead', depression == 'Having') %>%
                                  group_by(gender) %>%
                                  summarize(depressed_and_dead =length(depression)))
dead_depressed
gender <- merge(gender, dead_depressed, by = 'gender', all =
TRUE) # merging the result to gender table
barchart(depressed_and_dead~gender,
         data = dead_depressed,
         ylim = c(0,90),
         col = 'purple') ## show data value on the chart

### Question 6c Show the number of dead patients having or not having depression according to the genders

dep_not_dep<- as.data.frame(health_rec %>%
                              select(outcome, gender, depression) %>%
                              filter(outcome == 'Dead') %>%
                              group_by(gender, depression) %>%
                              summarize(num_of_patients = length(depression)))
## `summarise()` has grouped output by 'gender'. You can override using the
## `.groups` argument.
dep_not_dep
barchart(num_of_patients~depression|gender,
         data = dep_not_dep,
         group= depression,
         xlab = substitute(paste(bold('Depression'))),
         ylab = substitute(paste(bold('Number of patients'))))

Question 8

what is the rate of non-survived patients with hypertension?

dead_hyper <- health_rec %>%
  select(outcome, hypertensive) %>%
  filter(hypertensive == 'Having', outcome == 'Dead') %>%
  summarize(dead_hypertensive = length(hypertensive)) # this returns the patients that are dead and had hypertension

hypertensive <- health_rec %>%
  select(outcome, hypertensive) %>%
  filter(hypertensive == 'Having') %>%
  summarize(hypertensive = length(hypertensive)) #this returns all the patients with hypertension

dead_hyper$dead_hypertensive/  hypertensive$hypertensive * 100 #proportion of the dead hypertensive patients
## [1] 17.46988

Question 9

Rate of the different gender with hypertension per 100 patients

gender_hyper <- as.data.frame(health_rec %>% select(gender,
hypertensive) %>% filter (hypertensive == 'Having') %>%
group_by(gender) %>% summarize (hypertensive_tot =
length(hypertensive)))

gender <- merge(gender, gender_hyper, by = 'gender', all = T) #merging to gender
gender$rate_hyper <- gender$hypertensive_tot/gender$num * 100 ##estimating the rate
gender[, c('gender','rate_hyper')]

ANSWER 28 female and 29 male patients per 100 are having hypertension

###Question 9b patients having hypertension and dead

dead_gender_hyper <- as.data.frame(health_rec %>%
                                     select(outcome,gender, hypertensive)%>%
                                     filter (outcome == 'Dead',hypertensive == 'Having') %>%
                                     group_by(gender) %>%
                                     summarize (hypertensive_dead = length(hypertensive))) # this separates the dead hypertensive patients according to gender

ANSWER 31 female and 27 male of the dead patients had hypertension

gender <- merge(gender, dead_gender_hyper, by = 'gender', all = T)

Question 10

How many patients with renal failure are alive in the hospital?

health_rec %>%
  select(outcome, Renal_failure) %>%
  filter(outcome == 'Alive', Renal_failure == 'Having') %>%
  summarise(length(Renal_failure))

ANSWER 392 patients are having renal failure and alive

Question 10b

How many patients died having renal failure

health_rec %>%
  select(outcome, Renal_failure) %>%
  filter(outcome == 'Dead', Renal_failure == 'Having') %>%
  summarise(length(Renal_failure))

ANSWER37 patients died having renal failure

Question 10C1

People with renal failure according to gender

renal <- as.data.frame(health_rec %>%
     select(gender, Renal_failure)%>%
     filter(Renal_failure == 'Having') %>%
     group_by(gender) %>%
     summarise(renal_failure = length(Renal_failure)))

renal
gender <- merge(gender, renal, by = 'gender', all = T) #merging result to gender dataframe

10C2

num of patients that died having renal failure according to gender

renal_dead <- as.data.frame(health_rec %>% select(outcome,
gender,Renal_failure) %>% filter(outcome == 'Dead', Renal_failure ==
'Having') %>% group_by(gender) %>% summarise(length(Renal_failure)))
gender <- merge(gender, renal_dead, by = 'gender', all = T) #merging result to gender dataframe

Question 11

How many patients in the hospital with Hyperlipemia are dead?

dead_hyperli <- as.data.frame(health_rec %>%
                                select(outcome, gender, Hyperlipemia) %>%
                                filter(outcome == 'Dead', Hyperlipemia == 'Having') %>%
                                group_by(gender) %>%
                                summarize (dead_hyperli = length(Hyperlipemia)))

gender <- merge(gender, dead_hyperli, by = 'gender', all = T) # merging to gender

ANSWER rsum(gender$dead_hyperli) patients died having hyperlipermia

Question 11b

How many patients are having hyperlipemia

hyperli <- as.data.frame(health_rec %>%
                           select(gender, Hyperlipemia) %>%
                           filter(Hyperlipemia == 'Having') %>%
                           group_by(gender) %>% 
                           summarize(hyperlip = length(Hyperlipemia)))

729 patients are having hyperlipermia

gender <- merge(gender, hyperli, by = 'gender', all = T) # merging to gender

Question 12

how many patients in the hospital with Anemia are dead?

anemia <- as.data.frame(health_rec %>%
                          select(outcome, gender, deficiencyanemias) %>%
                          filter(outcome == 'Dead', deficiencyanemias == 'Having') %>%
                          group_by(gender) %>%
                          summarise(anemia_dead = length(deficiencyanemias)))
gender<- merge(gender, anemia, by = 'gender', all = T) # merging to gender

124 is the sum of dead patients having anemia

Question 12b

Total sum of patients having anemia

anemia_tot <- as.data.frame(health_rec %>%
        select(gender,deficiencyanemias) %>%
        filter(deficiencyanemias == 'Having') %>%
        group_by(gender) %>%
        summarise(anemia_tot = length(deficiencyanemias)))

gender <- merge(gender, anemia_tot, by = 'gender', all = T) #merging to gender

777 patients is having anemia

13

What is the proportion of survival and non-survival between diabetic and non-diabetic patients

Short preview of the stats relating to diabetic patients

health_rec %>% select(outcome, diabetes) %>% group_by(outcome,
diabetes) %>% summarize(count = length(diabetes))
## `summarise()` has grouped output by 'outcome'. You can override using the
## `.groups` argument.

13a1

The proportion of patients having diabetes

diabeteic <- as.data.frame(health_rec %>% 
                             select(diabetes) %>%
                             filter(diabetes == 'Having') %>% 
                             summarize(length(diabetes))) # this returns the total number of diabetic patients

diabeteic$`length(diabetes)`/length(health_rec$outcome) # this returns the percentage of diabetic patients
## [1] 0.5790816

ANSWER 57.9% of the patients are diabetic

13b

Proportion of diabetic that survived

To get the proportion of diabetic patient that survived, we need to know the total number of people having diabetic and we divide that by the number of people having diabetes and survived

having_diabetes_alive <- as.data.frame(health_rec %>%
   select(outcome, diabetes) %>%
   filter(outcome == 'Alive', diabetes == 'Having') %>%
   summarize(having_diabetes = length(diabetes))) # returns the patients having diabetes and still alive
  
having_diabetes <- as.data.frame(health_rec %>%
   select(diabetes) %>%
   filter(diabetes == 'Having') %>%
   summarize(diabetic = length(diabetes))) # returns patients having diabetes dead or alive

having_diabetes_alive$having_diabetes/having_diabetes$diabetic * 100 # proportion of diabetic patients alive
## [1] 85.02203

ANSWER 85% of diabetic patients are alive

13b2

Proportion of non-diabetic that survived. For this we follow the same procedure as the one above but this time, the patients are not diabetic

non_diabetic_alive <- health_rec %>% 
  select(outcome, diabetes) %>%
  filter(outcome == 'Alive', diabetes == 'Not Having') %>%
  summarize(non_diabetes = length(diabetes)) # this returns non-diabetic patients that are alive

non_diabetic <- health_rec %>%
  select(diabetes) %>%
  filter(diabetes == 'Not Having') %>%
  summarize(non_diabetic = length(diabetes)) #this returns non-diabetic dead or alive

non_diabetic_alive$non_diabetes/non_diabetic$non_diabetic * 100 # this returns the proportion of non-diabetic that are alive 
## [1] 88.48485

ANSWER88.4848485% of the patients are not having diabetes and are alive

13c

Proportion of diabetic that died

dead_diabetic <- health_rec %>%
  select(outcome, diabetes) %>%
  filter(diabetes == 'Having', outcome == 'Dead') %>%
  summarize (diabetic = length(diabetes)) # num of dead diabetic patients

dead_diabetic$diabetic/having_diabetes$diabetic * 100 #this returns the proportion of dead diabetics 
## [1] 14.97797

ANSWER about 14.98 percent of the diabetics are dead

13c2

Proportion of non_diabetic that died

dead_non_diabetic <- health_rec %>% 
  select(outcome, diabetes) %>%
  filter(diabetes == 'Not Having', outcome == 'Dead') %>%
  summarize (non_diabetic=length(diabetes)) # num of dead non-diabetic patients

dead_non_diabetic$non_diabetic/non_diabetic$non_diabetic * 100 # this returns the proportion of dead non-diabetics
## [1] 11.51515

ANSWER 11.52 percent of non diabetics are dead.

Question 14

What is the proportion of survival and non-survival between depressed and non-depressed patients

Short preview of the stats relating to depressed patients

health_rec %>% select(outcome, depression) %>%
  group_by(outcome, depression) %>%
  summarize(count = length(outcome))
## `summarise()` has grouped output by 'outcome'. You can override using the
## `.groups` argument.

Question 14a

What is the proportion of depressed and non-depressed that survived

depressed <- as.data.frame(health_rec %>% 
   select(depression) %>% 
   filter(depression == 'Having') %>%
   summarize(depressed = length(depression))) # this returns the total number of depressed patients either dead or alive

depressed$depressed/length(health_rec$outcome) # this returns the percentage of depressed patients
## [1] 0.8809524

ANSWER88.1 percent of the patients are depressed

14a1

Proportion of depressed patients that survived

having_depression_alive <- as.data.frame(health_rec %>% 
     select(outcome, depression) %>%
     filter(outcome == 'Alive', depression == 'Having') %>%
     summarize(having_depression = length(depression))) # returns the patients having depression and still alive

having_depression_alive$having_depression/depressed$depressed * 100 # proportion of depressed patients alive
## [1] 85.71429

ANSWER 85.7% of depressed patients are alive

14a2

Proportion of non-depressed that survived

For this we follow the same procedure as the one above but this time, the patients are not depressed

non_depressed_alive <- health_rec %>%
  select(outcome, depression) %>%
  filter(outcome == 'Alive', depression == 'Not Having') %>%
  summarize(non_depression = length(depression)) #this returns non-depressed patients that are alive

non_depressed <- health_rec %>% select(depression) %>%
filter(depression == 'Not Having') %>% summarize(non_depressed =
length(depression)) #this returns non-depressed patients count dead or alive

non_depressed_alive$non_depression/non_depressed$non_depressed * 100 # this returns the proportion of non-depressed patients that are alive
## [1] 92.14286

ANSWER92.1% percent of the patients are not having depression and are alive

14b1

Proportion of depressed patients that died

dead_depressed <-health_rec %>%
  select(outcome, depression) %>%
  filter(depression == 'Having', outcome == 'Dead') %>% 
  summarize (depressed = length(depression)) # num of dead depressed patients

dead_depressed$depressed/depressed$depressed * 100 #this returns the proportion of dead depressed patients
## [1] 14.28571

ANSWER about 14.28 percent of the depressed patients are dead

14b2

Proportion of non_depressed patients that died

dead_non_depressed <- health_rec %>% 
  select(outcome, depression) %>% 
  filter(depression== 'Not Having', outcome == 'Dead') %>% 
  summarize(non_depressed=length(depression)) # num of dead non-diabetic patients

dead_non_depressed$non_depressed/non_depressed$non_depressed * 100 # this returns the proportion of dead non-diabetics 
## [1] 7.857143

ANSWER 7.86 percent of the non-depressed patients are dead.

back to project page

portfolio page