A reflection on the application of grounded theory in the exploration of the experiences of informal carers

Kev words Informal carers; Hospital discharge; Elderly care; Community-based reha­ bilitation; Grounded theory Abstract: Curationis 32 (2): 21-30 The aim o f this paper is to reflect on the application o f a qualitative research method that presents novice researchers with a variety o f challenges. It is suggested that prospective users o f the grounded theory method should seek guidance from experts in the field. However, to find these experts has proved to be quite challenging. The research topic lends itself to a qualitative study in general using the grounded theory method in particular.


Introduction
The purpose o f this article is to dis cuss qualitative research in general with a particular focus on the applica tion o f the grounded theory method.
To place qualitative research in con text a brief overview o f research para digms is presented.
A paradigm is defined as a basic set o f beliefs that guide action (Denzin & Lin coln, 1998:185).Paradigms are human constructions and include three basic elements, i.e. ontology, epistemology and methodology.Ontology refers to the nature o f social reality, that is, how humans view the world.Denzin and Lincoln (2008: 11;1998:185) question, whether humans see the world as a sta ble place driven by laws (positivist view) or as a fluid place, constituted o f multiple social realities (interpretive view).They explain that epistemology gives an indication about the relation ship betw een researcher and the re searched.Positivists see this relation ship as objective and one that is based on the precise observations o f how things really work.While interpretives, hold the subjective view that during any investigation there is an interac tion betw een the researcher and the researched (Denzin & Lincoln, 2008: 10-M ethodology refers to how the re searchers w ould go about doing the research.The m ethodologies chosen by positivist researchers are experimen tal and m anipulative and aimed at the verification o f hypotheses.Interpretive methodologies are dialectical and aimed at the reconstruction o f previously held c o n stru c tio n s (D en zin & L in co ln , 1998:185).
It is clear that the choice o f m ethodol ogy would be based on the purpose o f the research and the researcher's view o f the nature o f hum an beings and so cial reality.R esearchers w ho adopt a p o sitiv ist approach deduce hypotheses from a general theory, gather data and use sta tistical m eans to test the theory's pre dictions (Neuman, 2003: 90).This ap proach implies the clinical observation o f human behaviour that is predictable and guided by laws o f cause and ef fect.Such an approach would not have served the purpose o f my research study that was exploratory in nature and aimed to describe human behav iour within a specific context.A ccord ing to McCann and Clark (2003:23) the classical grounded theory method de scribed by Glaser and Strauss (1967) took more o f a positivistic stance than that o f Strauss and Corbin (1990Corbin ( ,1998)).
Interpretive social scientists criticize positiv ists for not dealing w ith the m eanings that real people attach to ex periences; for not acknowledging that people have the capacity to think and reason and for ignoring that peoples' experiences are shaped by social con text (Neuman, 2003:76).
Interpretive social science is associ ated with the philosophy o f symbolic interactionism that embraces the con cept o f meaningful interaction (Neuman, 2003:90).The interpretive paradigm is based on the premise that human be ings create meaning in their worlds and that the m eaning is constructed as a result o f interaction with others.I used observation and dialogue to obtain re search data in my study, which implied a subjectivist relationship between par ticipants and myself.
From the above discussion o f research p arad ig m s, the ap p ro ach th a t best suited my research study was an inter pretive one.The interpretive paradigm assum es amongst others: • that scientific inquiry is aimed at the reconstruction o f pre viously held constructions • that the world is fluid in nature and the result o f multiple social realities • that during any investigation there is an interaction between the researcher and the re searched The positivistic paradigm, best suited for quantitative research studies differ fundamentally from that o f an interpre tive paradigm used in qualitative stud ies.Qualitative researchers have de scribed these differences (Denzin & Lincoln, 2008: 11;1998: 185).The dif ferences include the flexibility o f the research design, which means that in the case o f qualitative studies, the de sign unfolds as the fieldwork proceeds.This implies that the sample size is de term ined by research outcomes.
Qualitative research is mostly done in the field and the context in which the research is done is described in detail and co n stitu tes im portant research data.In phenomenological approaches the emphasis is on language while in the grounded theory method the em phasis is on context (Charmaz, 2006: 180;Lowenberg, 1993:57).Patton (2002: 40) identifies various themes for qualitative enquiry, which include, naturalistic enquiry, purposive theoretical sampling, direct personal experience and engagement, empathetic neutrality, context sensitivity and tri a n g u la te d en q u iry .T h ese th em e s served as a checklist to evaluate the appropriateness o f the research method used in this study.

The origin of grounded theory
G laser and S trauss d e v e lo p e d the grounded theory method in 1967.The theoretical orientation they held was that theories should be grounded in field data to reflect the actions, interac tions and social processes o f people.They stated that grounded theory is derived from data and then illustrated by characteristic examples o f data and that the usefulness o f such a theory lies in, amongst others, how it was gen erated, its clarity, density, scope, inte gration and fit (Glaser & Strauss, 1967: 5).
Glaser and Strauss, both sociologist, but from different schools o f thought, brought different perspectives to their publication The discovery o f Grounded theory in 1967.In their book they pre sented qualitative researchers with an alternative approach to theory genera tion.A theory that would fit empirical situations and would be understood by academics and lay-people alike.Glaser & Strauss described grounded theory research as, an initial, system atic dis covery o f the theory from the data o f social research (1967:3) Annells (1997a: 122), referred to as classic grounded theory research-crs.Glaser criticized Strauss and Corbin for proposing the paradigm model to establish links between categories, as a model could impact the emergence o f grounded theory.A second critique was levelled at the use o f the condi tional matrix as an analytical tool.Clas sic grounded theory is underpinned by symbolic interactionism, which looks at the individual in society.Strauss and Corbin (1990: 175) sug gested that the conditional matrix al lows the researcher to capture all the conditions and consequences that may impact the social interaction under in vestigation.The social interaction is explored beyond the individual in so c iety .In th e a p p lic a tio n o f th e grounded theory method, some o f the conditions and consequences are re lated to issues, like, e.g.national poli tics, religion and culture.McCann and Clark (2003:23) describe epistem ological and m ethodological differences between Glaser's (1967) and Strauss and C orbin's (1990, 1998) ap proaches to grounded theory.Accord ing to M cCann and Clark, Glaser fol lowed a more positivistic epistemology than Strauss and Corbin.Glaser also viewed the researcher's role as inde p en den t to th at o f the research ed , whereas Strauss and Corbin suggested an active and dialectic interaction be tw ee n th e re s e a rc h e r an d th e r e searched.M cCann and Clark further describe the methodological difference of, G laser's unassisted emergence o f theory during data analysis, to the ana lytical rules and procedures proposed by Strauss and Corbin.
These academic debates reinforced my decision to choose the approach de scribed by Strauss andCorbin (1990, 1998).The decision to use classic lit erature on grounded theory research is based on the fact that in recent pub lications, authors choose to discuss certain aspects o f the method, i.e. lit e ra tu re re v ie w in g an d re fle x iv ity (McGhee, Marland & Arkinson, 2007: 334), constructing grounded theory (C h a rm a z , 2 0 0 6 : 180), rig o r and grounded theory research (Chiovitti & Piran, 2003:427) and grounded theory in discipline specific research (McCann & Clark, 2003:23).The grounded theory research method as such, has not been sig n ific a n tly o r fu n d a m e n ta lly changed.

Grounded theory research -the method
Grounded theory research focuses on the generation o f theory with the em phasis on inductive strategies o f theory developm ent.Inductive analysis in volves the generation o f categories and themes from research data.Findings emerge through the researcher's inter action with the data.Sampling, data collection and analysis occur simulta neously (Patton, 2002:453).In contrast, by using deductive analysis the re searcher uses an existing framework to analyze data.Creswell (1998: 56) suggest that rigor ous enquiry, system atic procedures and specific methods underlie the gen eration o f grounded theory.He empha sizes that the theory should relate to the phenomenon (informal caring) be ing studied in its unique context, clarify the concepts or categories that emerge during the analysis o f data and indi cate the relationship between the con cepts.
Annells (1997b: 176) describe the es se n tia l elem e n ts in any g ro u n d ed theory study as amongst others, con stant comparative data analysis, theo retical sampling and sensitivity, theo retical saturation and the identification o f a core category.She traced these el ements to the original work o f Glaser and Strauss (1967) and Strauss and Corbin (1990).
The value o f grounded theory lies in its contribution to areas in which little research has been done (Charmaz, 2006: 17;Chenitz & Swanson, 1986:7)

Grounded theory -the method of choice
The grounded theory study that I have done at masters level entailed amongst others, purposive sampling, constant comparative data analysis but stopped short o f theory generation.During the preparatory reading for my doctoral studies, I discovered that grounded theory presents the researcher with more o f a challenge than an explora tory, d e sc rip tiv e q u alita tiv e study would.It challenged the researcher to m ove beyond the thematic analysis o f qualitative research data, to a level o f theory generation.
I have a fascination with the study o f the individual in society.I wondered about the impact o f prior experience on the burden o f care in an informal caring situation.The grounded theory method matched my academic interest about the experiences o f the individual in soci ety and academ ic aw areness o f the value o f inductive thinking processes.
I identified with the nurse researchers who expressed the value o f theory gen eration o f grounded theory research (McGhee et al, 2007;Chiviotti & Piran, 2003;Kendall, 1999;Annels, 1997b;Strauss & Corbin, 1990,1998).I could also appreciate the analytical tools pro vided by Strauss and Corbin (1990) to facilitate the im plem entation o f the method.
The grounded theory method provided the most appropriate way to explore the experiences o f in fo rm a l hom ebased c a re rs... across different social groupings.It allowed for the explora tion o f a research area in which little research has been done (Charmaz, 2006: 17;Chenitz & Swanson, 1986: 7).The grounded theory m ethod uses an in terpretive social research approach, which is generally used in exploratory social research.The clear philosophical underpinning and m ethodological guidelines pro posed by the founders and followers o f the grounded theory m ethod pro vided much needed structure to a nov ice researcher.I did, however, have a few false starts but having had to re do much o f the coding and conceptual analysis I had unwittingly strengthened my understanding o f the research proc ess.
S tudies about careg iv ers, inform al carers and fam ily carers have used grounded theory as a research method (R obinson, F rancis, Jam es, T indle, Greenwell & Rodgers, 2005:560;Shyu, 2000: 619).To join the academic dis cussion about the experiences o f these groups I decided to use sim ilar m eth ods o f enquiry.

Philosophical underpinning
Symbolic interactionism is the philo sophical underpinning o f grounded th e o ry re s e a rc h .S y m b o lic interactionism is both a theory about human behaviour and an approach to the investigation o f hum an conduct and group behaviour (Charmaz, 2006: 7;Annells, 1996:381).
G eorge Herbert Mead, a social psy chologist, postulated that it is in social interaction, that the individual achieves a sense o f s e lf (M ead, 1934).It is through interaction w ith others that one develops a sense o f one's roles in society.Herbert Blumer, a scholar o f M ea d , e la b o ra te d on sy m b o lic interactionism by identifying three ba sic prem ises to the philosophy.The first is that human beings act toward things on the basis o f what the things m ean to them.These things m ay be objects, symbols or the actions o f other human beings.The second is that the meaning o f such things is derived from social interaction with others.The last is that, following such an encounter, the person uses an interpretive proc ess to modify the meaning o f the things (Blumer, 1969:2).
T he a p p lic a tio n o f th e sy m b o lic interactionist perspective was dem on strated throughout my study in reflec tive and analytical memos.The inter pretation o f the participants' interac tions was based on their dialogue and the direct observation o f their behav iour within a specific context.

The application of grounded theory
The aim o f the study was to explore the experiences o f the informal carers o f elderly dependants in the M etropoli tan area o f Cape Town, South Africa.Linked to the exploration was the de velopment o f grounded theory on in formal care.The concept informal carer lacks a clear analytical definition.A c cording to Dahlberg, Demark & Bambra (2007:439) it centres on the care o f the elderly, people with chronic illness and disability and implies elements o f so cial or familial relationships.In this study, the informal carer is any person over the age o f 18 who assumes the primary responsibility to care for an eld erly person who, following hospital dis charge, requires assistance with two or more activities o f daily living.The eld erly dependants had to be discharged to a private dwelling and not to a nurs ing home, frail-care centre, hospice or hospital.
Sam pling, data collection and data analysis followed the tenets o f qualita tive enquiry.A small purposive sample o f eig h teen p a rtic ip a n ts (in fo rm al carers) was selected.Brink (2001:142) suggests a sample size o f 20 -30 sub jects for a qualitative study since too m any subjects could com plicate the data analysis process.The sample size was determined by a cyclical process o f data collection and data analysis and continued until the saturation o f data was achieved with eighteen informal carers.Saturation is the point at which no new information emerges from the interviews (Chiovitti & Piran, 2003:433).
Two tertiary hospitals in Cape Town were used as sites for data collection and potential carers were identified, using inclusive sampling criteria.
Q uestionnaires w ere used to collect dem ographic data from the informal carers and their elderly dependants, followed by interviews with the infor mal carers.The interviews, using openended questions, w ere scheduled as follows: prior to the discharge, within the first week o f the discharge and one month later at the homes o f the depend ants.The interview questions prior to hos pital discharge included, amongst oth ers: • Describe your feelings about the discharge o f X?
• W hy are you going to / have you chosen to take care o f X?
• Describe your needs / concerns with regard the discharge o f X?
• W hat changes have you made in preparation for the dis charge?The second and third interviews were scheduled at the home o f the elderly and included the following questions:

•
Describe your feelings about caring for X at home.The demographic data showed that the male and female carers ranged in age from 3 2 -7 5 years and that they repre sented all social groupings in the West ern Cape.Their relationships with the elderly were, amongst others: daugh ters or so n s, spo u ses or p a rtn e rs, grandchildren or siblings.M ale and female dependants were 65 -90 years old and they mostly suffered strokes, fem ur or hip fractures, and am puta tions.Field notes were used to capture additional contextual information about each informal carer.
In grounded theory research the chal lenge lies in the analysis o f the research d a ta to d e v e lo p a th e o ry th a t is grounded in the data.Creswell (1998: 57) and Strauss and Corbin (1990: 57;1998: 101) suggest a standard, system atic format o f data analysis in grounded theory research, i.e. open coding, axial coding and selective coding.

Open coding
Open coding means breaking down the data into smaller segments.It also in volves labeling these segments (con cepts), examining and comparing them and then grouping them under a more abstract concept (category).Catego  It is also important to employ the con sta n t co m p a ra tive m eth o d o f data analysis.This method allowed me to take the data collected and constantly compare it to existing categories.I then linked the data to emerging categories and compared the data at property and dimensional level (Strauss & Corbin, 1998: 85).At this level, I did not com pare individual cases but compared the categories.I was not interested in how m any cases displayed this particular category but how it presented itself under varying conditions (Strauss & Corbin, 1998:95).
During the initial open coding o f inter view transcripts, before using the com puter assisted data analysis software, the categories that emerged from the data related to choosing to care, pre paring to care, helping with care activi ties, challenges relating to caring, i.e. accessing assistive devices and state pensions and the need for respite care.
I also became aware that in m any in stances the caring became a shared re sponsibility.This happened in particu lar where the primary carer was male.
The assumption that females primarily accepted the role o f carer was proved incorrect during the early phase o f the data collection process.Despite hav ing daughters, two elderly ladies were cared for by their sons.
Having had the experience o f informal caring, I knew how it felt to be assigned to this role.The carer m ay choose to care or be chosen by others to provide informal home-based care to an elderly person.D uring the data collection phase, I purposively included both options, to ascertain w hether choos ing or being chosen, impacted on the care and rehabilitation offered to the e ld e rly d ep en d an t.I also b ecam e acutely aware o f my strong opinions w ith regard to the choosing to care .. .andcaptured these feelings in reflec tive notes.
The initial coding was followed by a more structured attempt to use com pu ter software to assist with the process.After immersing m yself in the data by reading and re-reading all the relevant documentation about a specific case I started open coding, using the N. VIVO programme.I consciously had to fo cus on the informal carers and their ex periences and found the seem ingly unstructured nature o f open coding difficult at first.The code segments varied in length from a few words to a paragraph.
The N.VIVO software facilitated the process o f open coding and tree cod ing.A line-by-line tree coding system was used and the software allowed for the retrieval o f codes as well as the link ing o f codes in and across cases.The challenge was to name the categories that would accurately describe the per ceptions, opinions or experiences o f the carers.I found the sim ilarities across cases interesting but also be came aware o f the variety o f responses to the same interview questions.Field notes and memos were linked to docu m ents and in-docum ent annotations were created.The com puter-assisted data analysis program m e was useful during the open coding phase o f data analysis.However, when conceptual linking o f concepts and categories was done, I preferred to use traditional meth ods o f data analysis and the computerassisted programme was abandoned.

Axial Coding
The next stage o f data analysis in v o lv e d s u g g e s tin g re la tio n s h ip s amongst categories, i.e. axial coding.Creswell (1998: 57) and Strauss and Corbin (1998:123) state that the data is now re-assembled in a conceptual way.C odes (nodes in the N .V IV O p ro gramme) were grouped together to form sub-categories.Sub-categories and categories were conceptualized in a lengthy constant comparison exercise.Sub-categories were linked to catego ries and eventually a basic social proc ess emerged.The aim o f axial coding is to identify a core category or basic so cial process -a central idea, event or happening in terms o f the paradigm model (Strauss & Corbin, 1990: 110).
The paradigm m odel is an analytical tool used to facilitate the exploration o f the core category or basic social process.By using the model I uncov- ered the core category namely, fa c ili tating care.It is by facilita tin g care that informal hom e-based carers make sense o f their w orld as they provide care and assist in the recovery o f their elderly dependants.W hen carers are faced with the prospect o f caring for elderly dependants, there are a num ber o f conditions that will impact on this basic social process.
Strauss and Corbin (1990) suggest that causal, contextual and intervening con ditions may influence the basic social process -BSP.They further suggest that the action and interaction strate gies used by participants engaged in the BSP may lead to specific conse quences.

Selective coding
Finally in selectiv e coding the re searcher identifies a "storyline" and writes a story that integrates the cat egories in the axial cod in g m odel (Creswell, 1998:57) Carers discover that the elderly may also require specialized repair care to ensure the recovery o f some ba sic functions, for example, they may need special exercises and special diets.
Carers becom e aware that caring takes its toll.They are faced with physical, emotional and social chal lenges.They become tired and frus trated at the lack o f support.Occa sionally family members will offer respite for short periods o f time.How ever support is disconnected and do not m eet the needs o f the carer at the time.In some cases the health services do not provide opti mal support and they have to rely on the community to assist with in formal home-based care.Chiovitti and Piran (2003:430) suggest that grounded theory researchers em ploy practical procedures to enhance standards o f rigor in their research.They define rigor as the credibility, auditability and fittingness o f the re search study.Credibility relates to the tru s tw o rth in e s s o f th e stu d y ; auditiability to the ability o f another researcher to follow the research proc ess o f the primary researcher and fit tingness or transferability to the likeli hood that the research findings will have meaning to others in the same situ ation.

Rigor and ethical consideration in grounded theory research
According to Chiovitti and Piran (2003: 430) the credibility o f the study is en hanced by letting the participants guide the research process and by using the actual words o f the participants in the theory.Initial codes were used to adapt the interview questions in order to fur ther explore these codes in subsequent interviews.During the second and third in terv iew sessio n s, the resea rch er checked and verified the provisional coding data with the participants, to refine the analysis (member checking).Anecdotes from interview transcripts were used to substantiate the research outcom es.M ore than one source o f evidence, i.e. questionnaires, inter views and field notes was used during the re se a rc h study.T ria n g u la tio n added to rigor and improved the cred ibility o f the data and an independent grounded theory m entor verified the data analysis process.
To account for auditablilty o f the re search study as described by Chiovitti and Piran (2003: 430) the researcher should make an effort to present a clear audit trail throughout the analysis proc ess.T his w as achieved by w riting transparent analytical and conceptual reasoning notes relating to the initial coding, the grouping o f codes into cat egories and the linking o f emerging categories to the basic social process.These notes formed part o f the research report.
To enhance fittingness in grounded theory research, it is important to iden tify the level o f theory generated.This research study was aimed at develop ing substantive theory that evolved from the exploration o f the phenom enon o f informal caring within a spe cific situational context (Strauss and Corbin, 1990:174).
The three ethical principles described by Brink (2001:39), i.e. respect for per sons, beneficence and justice, present researchers w ith clear guidelines to conduct and report the research study in an ethically sound way.
During the planning phase the proposal was presented and approved by the ethics committee o f the University.I applied for and was given permission by the Medical Superintendents o f the hospitals to conduct research within their health care facilities.
The potential carers were briefed about the aim o f the research study and in vited to ask questions for the purpose o f clarity.After explaining their rights, participants were asked to sign the let ter o f consent.Participation was vol untary and uncoerced and only one carer refused to participate in the re search study.
Permission was obtained to digitally record the interviews and the partici pants' names were used during these recordings.During the transcription o f the interviews the names were replaced initials (Mrs H) and numbers (Resp: 09).
A master list that provided the link be tween the names o f the participants and participant num bers was stored in a secure location.
It is important to m aintain anonymity during the reporting phase.References to race and gender w ere only done when it impacted on the interpretation o f the research outcomes.Participants were informed that the outcomes would be reported in, e.g. research publica tions and conference presentations.

Reporting grounded theory research
The biggest challenge in w riting a grounded theory report lies in the pres entation o f a substantive theory that reflects inductive and deductive rea soning, relationships between codes, categories and the basic social proc ess as well as the conditions that im pact on the basic social process.To achieve this balance the report should include anecdotes, field notes and memos.An excerpt related to role fit ting illustrates how the analysis proc ess was captured in the research re port:

Role fitting
Role fitting (see Figure : 2) represents the first strategy devised by the carer to manage and carry out the basic so cial process, i.e. facilitating care.Strauss and Corbin (1990: 97) state that action strategies occur under a certain set o f perceived conditions.This implies that the more compromised the elderly, the m ore intensive, facilitating care be comes.They suggest that all action strategies should be discussed in terms o f their properties namely process, pur pose and goal.Process refers to a timeline, i.e. when does role fitting start?Researchers also have to specify the purpose and outcom es o f fitting the role o f facilitator o f care into existing roles.The insertion o f analytical notes makes the cognitive processes overt, e.g.A nalytical notes: In grounded theory it is important to indicate the relationship between the categories and indicate the relationship between the different concepts.It is also important to discuss the variations within the cat egory (Strauss & Corbin, 1990:253).In the case o f role fitting, there is a close interplay between prior relationship, the incident and the current roles o f the prospective carer.
The purpose o f role fitting is analyzed, i.e. it allows prospective carers the time to make sense o f the new role.The goal o f role fitting is to accept the role, to explore the expectations o f the role and possibly to prepare for the role.It al lows the facilitator o f care an opportu nity to obtain inform ation, develop skills and access appropriate assistive devises to perform the role.
The conditions that impact on the cat egory role fittin g are explored.In this study the success o f role fitting depend on the following sets o f perceived con ditions: • the carer chooses to care • the carer has experience • the carer prepares fo r the role

• the carer is the sole provider o f care
The carer chooses to care is analyzed in single cases and across cases.Most carers who chose to facilitate care had an established prior relationship with the elderly.A wife chose to care for her husband after fifty years o f married life.
A gentleman chose to care for a part ner with whom he had an adulterous prior relationship.In the cases where the carer chose to care a conscious ef fort was made to fit the role into their existing roles.Role strain and role con flict developed when carers were cho sen by others to facilitate care.A young mother was asked by her family to move back to her parent's house after her mother suffered a stroke.Carers cho sen by others may experience difficulty (perceived or real) in fitting the role o f facilitator o f care into existing roles.
Reflective notes: Health care workers should become more sensitive to the concept o f role fitting and how it im pacts on facilitating care.
A nalytical notes: Prior relationships clearly have an effect on role fitting.Carers are occasionally chosen on the basis o f a prior parent-child relation ship without taking the adult child's current portfolio o f roles into account.
A consequence could be disconnected care resulting in severed social relation ships.
All the perceived conditions are ex plored in this fashion.This process continues until all the categories reach saturation.

Discussion
Qualitative research in general and the grounded theory m ethod in particular, presents novice researchers with a va riety o f challenges.It is therefore im portant to self-assess and to reflect on the appropriateness o f the m ethodol ogy.To this end I have used an evalu ation checklist proposed by Patton (2002: 40).The evaluation is based on the following themes: naturalistic en quiry, purposive theoretical sampling, direct personal experience and engage m ent, em pathetic neutrality, context sensitivity and triangulated enquiry.
Informal hom e-based caring occurs in the real world and was thus not m a nipulated or controlled for the purpose o f the enquiry (naturalistic enquiry).
Since observations o f informal carers occurred in real-world settings, inter views were scheduled either in the hos pital ward in the home where carers were facilitatin g care.
A ccording to Patton (2002: 40) the qualitative researcher should be able to do purposive, theoretical sampling.It was, therefore, not im portant to con trol variables but rather to ensure that data was explored to its fullest.The qualitative data that I have gathered is thick in its description o f the informal hom e-based carer, the elderly as well as the setting in which the caring oc curred.W hen the category p rio r rela tionship was first coded I purposefully searched for data to explore its dimen sions.
Direct contact was sought with infor mal hom e-based carers (direct p e r so n a l experience an d engagem ent) w h o se p e rs o n a l e x p e rie n c e s and insights became part of, and critical to the understanding o f the phenomenon o f informal caring.During the interviews a conscious attempt was made to dis play em pathetic neutrality (Patton, 2002: 40).This meant that informal carers were shown respect, responsive ness and understanding.Being nonjudgm ental facilitated an environment where participants were willing to share information that allow ed an insider's perspective o f their realities.
Informal caring is a dynamic process that m ay change during the period o f enquiry.The strategies used during the data analysis process reflected this dynamism.Each carer's reality was dif ferent and the detail o f each case was review ed before cross case analysis was performed.Emerging categories were explored, themes identified and interrelationships confirmed by analyti cal principles rather than statistical rules. For

Limitations of the research study
A small-scale study has its limitations in that the sample size may limit the transferability o f the outcomes.How ever, even though the sample size may be considered a limitation, the depth and insight gained from such a study cannot be replicated by using quanti tative methods.
Another limitation is common to many qualitative studies.Qualitative studies are not aimed at presenting research outcom es that are generalizable but rather, outcomes that represent the re alities o f respondents who provide in formal care within their own unique contexts.
The study focused on the period o f transition o f the elderly from hospital to home, and in particular during the first m onth follow ing hospital d is charge.Even though the study is lim ited to the early stages o f caregiving im portant base-line information was collected.Ideally, a study that explores informal caring after a period o f one year w ould be an appropriate follow -up study.

Research outcomes and recommendations
The research outcom es show ed that informal carers were facilitating care during the transition o f the elderly from hospital to home by revealing the link betw een facilitating care, the basic social process, and other categories associated with informal health care.These categories include: the prio r re lationship between the carer and the elderly, the traumatic incident, the need for role fitting, maintenance-and re p air care, as well as, the consequences o f facilitating care, i.e. connected or disconnected care.
There is no official recognition o f the value o f inform al carers w ithin the health care structures in this country.
The incorporation o f carers into the District Health System would indirectly ease the burden o f care from the formal health care services.It is recommended that policy docum ents be developed to address the rights o f the carers and to recognize their value as facilitators o f care within communities.In particu lar, the right to be adequately prepared to provide care and rehabilitation serv ices within the home.
The shift in healthcare provision from curative, institutionalized services to wards com m unity-and hom e-based services required that additional re sources be allocated to support these services.It is recommended that the re allocation o f resources be prioritized by the Department o f Health.
There seem to be gaps in the referral system between the tertiary hospitals and community health care facilities.
Research, to explore these gaps in the referral system within the health sys tem in the Western Cape, needs to be done as a m atter o f urgency.
It is also recom m ended that a suitably trained person be appointed to act as liaison officer betw een the hospital staff, community-based personnel and home-based carers.Informal carers ex pressed the need for additional infor mation about the dimensions o f care and guidelines to access resources.

Conclusion
The The implementation phase was chal lenging at many levels.The process was daunting at first, with the labelling o f codes and categories and uncover ing the basic social process.However, this prolonged engagem ent with the research data allows for the develop m ent o f a substantive theory that is truly grounded in the data.
The author wishes to extend a sincere thanks to the carers who participated in the study and the National Research F oundation (N R F) w ho funded the project.

References
ries are further analyzed by exploring th e ir p ro p e rtie s an d d im e n sio n s (Strauss and Corbin 1998:101).

G
laser (in Melia 1996: 368)  describes the process o f segmenting the data as distracting.I agree to a certain extent, namely, that if too many small segments o f data are explored, one may loose the sense o f what is really going on in the data.I found that by constantly re flectin g on the ex p erien ces o f the carers in a particular context facili tated the process o f open coding.M y research mentor assisted with veri fication o f the process o f data analysis .She expressed concern about assign ing in-vivo codes to the feelings ex pressed by the participants.I had to account for these in-vivo codes by demonstrating how the code-category links were made, e.g.linking eveiything was topsy-turvy and things were hec tic to the category dis-connected care.The following serves as examples o f open coding: (See Figure: 1) Initial codes (6 o f 113 codes) Carer chooses to care Carer chosen Carer experienced Carer novice Carer shares caring Carer prepares for caring These codes were grouped along with others into the category: Role fittin g (1 o f 6 categories) Categories (e.g.Role fitting) were ex plored in terms their properties, e.g. the goal and purpose o f the category.Fur thermore the processes within a cat egory are discussed in terms o f tim e lines, e.g. when does it occur and does it change over a period o f time.Strauss and Corbin (1998: 101) further suggest that the dimensions o f the cat egory (Role fitting) be explored, i.e. the variation within the properties o f the specific category, i.e.Carer prepares -carer unprepared Carer experienced -carer novice

Figure
Figure 1: Example of Open Coding

Figure: 2 -
Figure: 2 -Role Fitting (Action and interaction strategies) E L L S , M 1997b: G rounded theory method, part II -Options for users o f the method.Nursing Inquiry, 4,76-180.B L U M E R , H 1 9 6 9 : S y m b o lic in teractio n ism .E ng lew o o d C liffs: Prentice Hall B R IN K , H 2001: Fundamentals o f re search m ethodology for health care professionals.Cape Town: Juta.C H A R M A Z , K 2006: Constructing G rounded theory: A practical guide through Q ualitative analysis.Thou sand oaks: Sage Publications C H E N IT Z , W C & SW ANSON, JM 1986: From p ra c tic e to g ro u n d ed theory: qualitative research in nursing.M enlo P ark, C alifo rn ia: A d dison-Wesley.C H IO V IT T I, R F & P1RAN, N 2003: Rigour and grounded theory research.J o u rn a l o f A d v a n c e d N u rsin g .44{4):427-435.C R E S W E L L , JW 1998: Qualitative inquiry and design: choosing among five traditions.Thousand Oaks: Sage Publications.D A H L B E R G , L ; D E M A C K , S & BA M BRA , C 2007: Age and gender o f informal carers: a population-based study in the UK.Health and social care in the Community.15(5^:439-445.D EN ZIN ,N K & LINCOLN, YS 2008: Collecting and Interpreting qualitative materials.Thousand Oaks: Sage Publi cations DENZIN, NK & LINCOLN, YS 1998: Strategies o f qualitative inquiry.Thou sand Oaks: Sage Publications GLASER, B G & STRA U SS, AL 1967: The discovery o f grounded theory: strategies for qualitative research, Chi cago: Aldine.GUBA, E & LINCOLN, Y 1985: Effec tive evaluation: im proving the useful ness evaluation results through re sponses and natu ralist approaches, San Francisco: Jossey-Bass.K EN D A LL, J 1999: Axial coding and G rounded theory controversy.W est ern Jo u rn al o f N u rsin g R esearch .21(6):743-757.LO W EN BERG , JS 1993: Interpretive research methodology: broadening the dialogue.A dvances in N ursing Sci ence.16(2):57-69.M EA D , G H 1934: Mind, Self and So ciety.Chicago: University o A N N , T & C L A R K , E 2003: Grounded theory in N ursing research: part 2 -Critique.Nurse Researcher.11(2): 19-28.M cG H E E , G ; M A R L A N D , G R & ATKINSON, J 2007: Grounded theory research: literature reviewing and reflexivity.Journal o f A dvanced Nursing.60(3):334-342.M E L IA , K M 1996: Rediscovering Glaser.Qualitative Health Research.6(3):368-378.NEUM AN, W L 2003: Social research methods: qualitative and quantitative approaches.Boston: Allyn and Bacon.PATTON, M Q 2002: Qualitative re search & evaluation methods.(Third Edition) California: Sage Publications.ROBINSON, L; FRANCIS, J; JAM ES, P; TIN D L E , N; G R E E N W E L L , K & RO D G ER S, H 2005: Caring for carers o f people w ith stroke: developing a com plex intervention follow ing the Medical Research Council framework.Clinical Rehabilitation.2005.19:560-571.SHYU,Y-IL 2000: The needs o f family caregivers o f frail elders during the tran sition from hospital to home: a Taiwan ese sample.Journal o f Advanced Nurs ing.32(3 ):619-625.ST E R N , PN 1994: Eroding grounded theory.In Morse.J. Editor.Critical is sues in qualitative research methods.Thousand oaks: Sage.STRA U SS, A & C O R B IN , J 1990: B a sic s o f q u a lita tiv e re s e a rc h : grounded theory procedures and tech niques.California: Sage Publications.STRA U SS, A & C O R B IN , J 1998: Basics o f qualitative research: tech niques and procedures for developing gro u n d e d th eo ry .T h o u sa n d oaks: Sage Publications.